T.C.
Mersin Oniversitesi
Sosyal Bilimler Enstitiisii
ingiliz Dili ve Edebiyat1 Anabilim Dah
14 ?-J 2
PSYCH VERBS IN TURKISH: A STRUCTURAL AND SEMANTIC ANALYSIS
Ptnar iBE
Dam~man
Dos;. Dr. Mustafa AKSAN
YUKSEK LiSANS TEZi
Mersin, 2004
Mersin Universitesi Sosyal Bilimler Enstitusu MudurlUgline,
Bu yah~ma jurimiz tarafmdan ingiliz Dili ve Edebiyati Anabilim Dah 'nda
YOKSEK LiSANS TEZi olarak kabul edilmi~tir.
Ba~kan _ _ _r::J_r~-----Do9. Dr.
Ye~im
AKSAN
Doy. Dr. Omit Deniz TURAN
Dye~
:.nr:MU:fuAK.sAN
(Dan1~man)
Onay
Yukandaki imzalarm ad1 geyen ogretim elemanlarma ait olduklanm onaylanm.
~-
- ------~
c~ - r-·~-r
Prof. Dr. Serra DURUGONUL
Enstitu Mud1irii
08/07/2004
11
OZET
Bu c;ah~ma, konu rolleri ve sozdizim konumlan arasmda duzenli bir ortii~menin
varhgmt onvarsayan baglama/
ortu~me
kuramlanna evrensel bir bic;imde aykm
davrant~
gosteren bir grup eylem olan ruh durumu eylemlerinin Ttirkc;edeki gorunumlerinin
betimsel bir c;ozumlemesini sunar ve bu eylemler ic;in ozgun bir smif1andtrma onerir.
<;ah~manm
gurubunun, bu
anlama
ilk boltimu,
davranl~lnl
ili~kin yakla~tmlan
<;ah~manm
suad1~1
ac;tklamaya
davrant~
c;ah~an,
gosteren bu ey lemlerden bir
farkh olc;utlere
dayandmlmt~
yapt ve
ozetler.
ikinci boltimu, Turkc;e ruh durumu eylemlerini, uyelerinin konu
rolleri, durum ekleri ve sozdizim konumlarmt temel alan, onerilen bir smlf1andtrma
c;erc;evesinde inceler. Bu stmflandtrmaya gore, Turkc;ede, ilk ikisi
durum ekli
gec;i~siz
gec;i~li,
digerleri farkh
dort gurup ruh durumu eyleminden soz edilebilir. Sorunlu ikinci gurup
eylemler, gozlemlenen evrensel egilimlere benzerlik gosterir. Yalmzca Deneyimleyen
uyesine sahip dorduncu gurup eylemler ise, alanyazmda ruh durumu eylemleri
incelenmi~
dillerin hic;birinde yer almamalan sebebiyle Turkc;eye ozgu gorunmektedir.
<;ah~ma
Turkc;ede son derece verimli bir bic;imde ruh durumu bildirme
i~levi
ustlenen bile~ik yaptlan, anlambilim ve durum temelli bir yakla~tmla inceler.
<;ah~manm
son boltimu ise ruh durumu bildiren stfatlann ttiretim ozelliklerini
bi<;im ve anlam yonunden ele ahr.
111
Tiirk9edeki ruh durumu eylemlerinin ettirgen, edilgen,
yapllarda nas1l bir
davram~
sergiledigine
onerilen bu smlflandmna olu~turur.
ili~kin
tiirn
donii~lii, i~te~
tartl~malarm
ve
bile~ik
temelini sozi.i edilen
IV
ABSTRACT
This study gives a descriptive account of psychological state verbs in Turkish
which seem to be universally problematic for linking/ mapping theories that assume a
unified mapping between the thematic roles and the syntactic positions; proposing an
original classification.
The first section of the study summarizes the structural and semantic
approaches which attempt to explain the unexpected behavior of a group of psych verbs
based on different criteria.
In the second section, Turkish psych verbs are analyzed through the proposed
classification based on the thematic roles, the case marking and the syntactic position of
the arguments. According to the classification, there are four groups of psych verbs in
Turkish; the first and the second are transitive and the others are intransitive with different
case marked Theme objects. The problematic second group conforms to the observed
universal tendencies. The fourth group with an only Experiencer argument; on the other
hand, seems to be the only group which does not exist in other languages in the literature
whose psych verbs are analyzed.
The compound forms which seem to be a very productive way of conveying
psych senses in Turkish are also examined with a semantic and case marking based
approach.
v
Lastly, the derivational properties of the psych adjectives are also examined
both morphologically and semantically.
The mentioned proposed classification constitutes the basis of all discussions
analyzing the behavior of Turkish psych verbs with respect to causativity, passivization,
reflexivity, reciprocity and compound forms.
VI
CONTENTS
ACKNOWLEDGEMENT ........................................................................... i
OZET ................................................................................................... .ii
ABSTRACT ........................................................................................... .iv
CONTENTS .............. ............................................................................. vi
THE LIST OF TABLES ...................................... .. ............ . ..... ........ ........... xi
THE LIST OF FIGURES ................................................... ...... .............. xviii
INTRODUCTION ..................................................................................... 1
Background Information .............................................................................. 1
The Purpose of the Study .............................................................................. 8
Research Questions ......................................................................... .. ......... 9
Hypotheses ............................................................................................. 9
Data Collection Techniques and Limitations ..................................................... 10
The Method of Analysis ............................................................................. 10
I.
REVIEW OF LITERATURE ..................................... ...................... 12
I. 1.
Argument Structure ....................................................................... 12
I. 1. 1. Transitivity and the Unaccusative Phenomenon ................................ 23
I. I. 2. Voice Alternations and Causativity .............................................. 27
I. 2.
Approaches to Psych Verbs .............................................................. 33
I. 2. 1. Structural Approaches to Psych Verbs ................................................ 34
I. 2. 1. 1.
Belletti and Rizzi ( 1998) and the Psych Verbs in Italian ............. 34
I. 2. I. 2.
Pesetsky ( 1987) and the Binding Problems with Experiencer
Verbs ..................................................................................................... 42
Vll
I. 2. 1. 3.
Vanhoe (2002) and the Psych Verbs in Spanish ...................... .47
I. 2. 1. 4.
Mulder (1992) and the Unergative and Ergative Psych Verbs ..... .48
I. 2. 1. 5.
Hale and Keyser (2002) on Psych Verbs .............................. .49
I. 2. 1. 6.
Ginnis (nd.) on the Morphological Restrictions in Experiencer
Predicates .............................................................................................. 51
I. 2. 1. 7.
Kural (1996) on the Elementary Predicates and the Psych Verbs in
Turkish ...................................... , .......................................................... 52
I. 2. 2. Semantic and Aspectual Approaches to Psych Verbs ............................... 57
I. 2. 2. 1.
Grimshaw (1990) Prominence Theory and Psych Verbs ............. 58
I. 2. 2. 2.
Tenny (1994) Aspectual Interface Hypothesis and Psych Verbs .... 59
I. 2. 2. 3.
Voorst ( 1992) and the Aspectual Semantics of Psych Verbs ........ 63
I. 2. 2. 4.
Filip (1996) on the Psych Verbs in Czech .................... ; ......... 66
I. 2. 2. 5.
Arad (nd.) on Psych Verbs ................................................ 68
I. 2. 2. 6.
Levin and Hovav (2002) on Psych Verbs .............................. 70
I. 2. 2. 7.
Dowty (1991) on Proto Roles and Psych Verbs ....................... 71
I. 2. 2. 8.
Hatory (1997) and the LCSs of Psych Verbs .......................... 72
I. 2. 2. 9.
Wechsler ( 199 5) Notion Rule and Psych Verbs ....................... 75
I. 2. 2. 10.
Krifka (2001) on Psych Verbs ........................................... 77
I. 2. 2. 11.
Kordoni (2000 and 2001) on the Psych Verbs in Modem Greek ... 79
I. 2. 2. 12.
Dabrowska (1996) on the Dative and Nominative Experiencers in
Polish ................................................................................................... 82
II.
ANALYSIS OF THE PSYCH VERBS IN TURKISH .............................. 84
II.
1.
II. 1. 1.
What is Psych? ..................................................................... 84
The Experiencer Argument ............................................... 86
viii
II. 1. 2.
The Theme Argument ..................................................... 87
II. 1. 3.
The Input of Psych Verbalization ....................................... 88
II. 1. 4.
Secondary Psych Senses .................................................. 93
II. 2.
Psych Verb Types in Turkish .................................................... 94
II. 2. 1.
Type 1 Psych Verbs ....................................................... 96
II. 2. 2.
Type 2 Psych Verbs ....................................................... 98
II. 2. 3.
Type 3 Psych Verbs ...................................................... 100
II. 2. 3. 1.
Type 3a Psych Verbs .......................................................... 100
II. 2. 3. 2.
Type 3b Psych Verbs ......................................................... 102
Type 4 Psych Verbs ....................................................................... 107
II. 2. 4.
II. 3. Voice and Psych Verbs in Turkish ..................................................................... 109
II. 3. 1.
Psych Verbs and Transitivity in Turkish .................. 110
II. 3. 2.
Psych Verbs and Causativity in Turkish .................. 112
~L
II. 3. 2. 1.
Morphologically Causative Psych Verbs ................. .J 14
II. 3. 2. 1. 1.
Type 1 Psych Verbs and Morphological Causativity ...... .114
II. 3. 2. 1. 2.
Type 2 Psych Verbs and Morphological Causativity ....... 122
II. 3. 2. 1. 3.
Type 3a Psych Verbs and Morphological Causativity ...... 126
II. 3. 2. 1. 4.
Type 3b Psych Verbs and Morphological Causativity ...... 129
II. 3. 2. 1. 5.
Type 4 Psych Verbs and Morphological Causativity ....... 133
II. 3. 2. 2.
Periphrastic Causative Psych Verbs ........................ 136
II. 3. 2. 2. 1.
Type 1 Psych Verbs and Periphrastic Causativity .......... 136
II. 3. 2. 2. 2.
Type 2 Psych Verbs and Periphrastic Causativity .......... 139
II. 3. 2. 2. 3.
Type 3a Psych Verbs and Periphrastic Causativity ......... 141
II. 3. 2. 2. 4.
Type 3b Psych Verbs and Periphrastic Causativity ......... 142
lX
II. 3. 2. 2. 5.
II. 3. 2. 3.
Causativity and the Permissive Reading .................. 144
II. 3. 2. 4.
Control, Causativity and the Psych Events ............... 146
II. 3. 3.
Psych Verbs and Passivization in Turkish ........................ .150
II. 3. 3. 1.
Type 1 Psych Verbs and Passivization ....................... .151
II. 3. 3. 2.
Type 2 Psych Verbs and Passivization ........................ 153
II. 3. 3. 3.
Type 3a Psych Verbs and Passivization ...................... 156
II. 3. 3. 4.
Type 3b Psych Verbs and Passivization ..................... .158
II. 3. 3. 5.
Type 4 Psych Verbs and Passivization ........................ 159
II. 3. 4.
Psych Verbs and Reflexivity in Turkish ............................ 160
II. 3. 4. 1.
Type 1 Psych Verbs and Reflexivity ......................... .160
II. 3. 4. 2.
Type 2 Psych Verbs and Reflexivity .......................... .161
II. 3. 4. 3.
Type 3a Psych Verbs and Reflexivity ........................ 164
II. 3. 4. 4.
Type 3b Psych Verbs and Reflexivity ........................ 164
II. 3. 4. 5.
Type 4 Psych Verbs and Reflexivity ......................... .165
II. 3. 5.
II. 4.
Type 4 Psych Verbs and Periphrastic Causativity .......... 143
Psych Verbs and Reciprocity ......................................... 166
Compound Psych Use-sin Turkish .... ; ............................................... .167
II. 4. 1.
Compound Psych Verbs which Derive with a Nominative Psych
Nominal .............................................................................................. 169
II. 4. 2.
Compound Psych Verbs which Derive with an Accusative Psych
Nominal .............................................................................................. 175
II. 4. 3.
Compound Psych Verbs which Derive with a Dative Psych
Nominal ............................................................................................. 177
X
II. 4. 4.
Compound Psych Verbs which Derive with a Complex Nucleus
......................................................................................................... 180
II. 4. 5.
II. 5.
Compound Psych Verbs which Derive with an Adjective .......... 180
Psych Adjective Derivation .................................................... 184
II. 5. 1.
Psych Adjectives Derived from Verbal Roots ........................ 185
II. 5. 2.
Psych Adjectives Derived from Adjectival Roots ................... 191
II. 5. 3.
Psych Adjectives Derived from Nominal Roots ..................... 191
II. 5. 4.
Non- Derived Psych Adjectives ........................................ 194
II. 5. 5.
Compound Psych Adjectives ............................................ 195
CONCLUSION .................................................................................. ... 197
BIBLIOGRAPHY .................................................................................. 202
XI
THE LIST OFTABLES
I. 1.
Table 1.
Dowty (1991) Thematic Roles and the Syntactic Realization Rules ......... 19
1.1.1.
Table 1.
Syntactic Accusativity ............................................................ .25
Table 2.
Syntactic Ergativity ................................................................ 25
Table 3.
Nakipoglu (1998) Unaccusatives and Unergatives ............................. 27
I. 2.
Table 1.
Belletti and Rizzi (1988) Psych Verbs in Italian .............................. .41
Table 2.
Vanhoe (2002) Psych Verbs in Spanish ........................................ .48
Table 3.
Mulder (1992) s- psych verbs and o-psych verbs ............................. .48
Table 4.
Kural (1996) Affectedness and Protogonism .................................... 53
Table 5.
Kural (1996) Affectedness, Protogonism and Thematic Roles ............... 53
Table 6.
Kural (1996) Thematic Combinations with the Experiencer Argument .... 54
Table 7.
Dabrowska (1994) The two Aspects of Mental Experiences .................. 82
Table 8.
Dabrowska (1994) Nominative and Dative Experiencers ..................... 82
II. 1.
Table 1.
Psych Noun+ -!An= Psych Verb ................................................. 88
Table 2.
Psych! Non-Psych Noun+ -!A= Psych Verb .................................... 89
Table 3.
Verb-> Psych Noun-> Psych Verb ............................................. 89
Table 4.
Verb -> Psych Adjective -> Psych Verb ......................................... 90
Table 5.
Verb+ Seemingly Causative Morpheme= Psych Verb .......................... 90
Table 6.
Verb+ Seemingly Passive Morpheme= Psych Verb ............................ 90
Table 7.
Verb+ In! 1$= Psych Verb ......................................................... 91
xu
lA~=
Table 8.
Psych Adjective+
Psych Verb .............................................. 91
Table 9.
Psych/ Non-Psych Adjective+ -(!An)= Psych Verb ........................... 92
Table 10.
Psych/ Non-Psych Adjective+ -sA= Psych Verb ............................... 92
Table 11.
Non-Psych Adjective+ -Arl AI= Psych Verb ................................... 92
Table 12.
Verbs which have both psych and non psych senses ........................... 94
II. 2.
Table 1.
Psych Verb Types in Turkish ..................................................... 96
II. 3.
Table 1.
Voice in Turkish .................................................................. 109
Table 2.
Tietze ( 1989) Medioreflexive Verbs ........................................... 110
II. 3. 1.
Table 1.
Psych Verbs and Other Intransitives ............................................ 110
Table 2.
Transitivity in Turkish Psych Verbs ............................................ 110
Table 3.
Optional Argument Deletion of Base Type 3a Verbs ........................ 111
Table 4.
Optional Argument Deletion of Derived Type 3a Verbs .................... 111
Table 5.
Optional Argument Deletion of Type 3b Verbs .............................. 111
II. 3. 2. 1.
Table 1.
Type 1 Psych Verbs and Causativity ............. ; ............................. 114
Table 2.
Type 1 Causative Psych Construction .......................................... 116
Table 3.
Type 1 Causative Psych Construction (Revised) .............................. 120
Table 4.
Type 2 Psych Verbs and Causativity .......................................... .122
Table 5.
Type 2 Causative Psych Construction .......................................... 125
Table 6.
Type 3a Psych Verbs and Causativity ......................................... 126
Table 7.
Type 1 and Type 3a Causative Psych Construction (Comparison) ......... 126
Xlll
Table 8.
Type 1 and Type 3a Causative Psych Construction (Differences) .......... 127
Table 9.
Type Type 3a Causative Psych Construction .................................. 128
Table 10.
Type 3b Psych Verbs and Causativity .......................................... 130
Table 11.
Type 3b Causative Psych Construction ........................................ 132
Table 12.
Type 4 Psych Verbs and Causativity ........................................... 133
Table 13.
Type 4 Causative Psych Construction .......................................... 134
Table 14.
Psych Verbs which Behave like the Lexical Causatives of Type 2 ........ 135
II. 3. 3.
Table 1.
Passivization in Psych Verbs ..................................................... 150
Table 2.
Type 1 Psych Verbs with -II ..................................................... 151
Table 3.
Type 1 Psych Verbs with -(l)n .................................................. 152
Table 4.
Type 1 Passive non psych Verbs ................................................ 152
Table 5.
Passive and Active Psych Verbs ................................................ 152
Table 6.
Active Type 2- Passive Type 3a Verbs ......................................... 156
Table 7.
Type 3b Verbs with Passive Morphemes with a Reflexive Meaning ...... 158
Table 8.
Type 4 Passive Verbs with a Reflexive Meaning ............................. 159
II. 3. 4.
Table 1.
Some Type 1 Verbs and Reflexivity ............................................ 160
Table 2.
Type 2 Verbs and Reflexivity ................................................... 162
Table 3.
Seemingly Reflexive Psych Verbs ............................................. 166
II. 4.
Table 2.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (2)
.......................................................................................................... 168
Table 3.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (3)
.......................................................................................................... 169
XIV
Table 4.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (4)
.......................................................................................................... 169
Table 5.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (5)
.......................................................................................................... 170
Table 6.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (6)
........................................................................................................... 170
Table 7.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (7)
........................................................................................................... 171
Table 8.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (8)
......................................................................................................... .171
Table 9.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (9)
.......................................................................................................... 172
Table 10.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (10)
......................................................................................................... 172
Table 11.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (11)
... . ························· .................... ··············· ··························· .............. 173
Table 12.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (12)
.......... . .............................................................................................. 174
Table 13.
Compound Psych Verbs Which Derive with a Nom. Psych Nominal (13)
.......................................................................................................... 174
Table 14.
Compound Psych Verbs Which Derive with an Ace. Psych Nominal (1)
.......................................................................................................... 175
Table 15.
Compound Psych Verbs Which Derive with an Ace. Psych Nominal (2)
....................................................................................... ....... ... . ........ 176
Table 16 .
Compound Psych Verbs Which Derive with an Ace. Psych Nominal (3)
..................................................................................... . .................... 176
Table 17.
Compound Psych Verbs Which Derive with an Ace. Psych Nominal (4)
.......................................................................................................... 176
Table 18.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (1)
. ........................................................................................................ 177
Table 19.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (2)
..................................... ······ ·········· ................. ················ ······ ............. 177
XV
Table 20.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (3)
......................................................................................................... 177
Table 21.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (4)
......................................................................................................... 178
Table 22.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (5)
......................................................................................................... 178
Table 23.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (6)
... ........ .. ............................................................................................ 179
Table 24 .
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (7)
......................................................................................................... 179
Table 25.
Compound Psych Verbs Which Derive with a Dat. Psych Nominal (8)
......................................................................................................... 179
Table 26.
Compound Psych Verbs Which Derive with a Complex Nucleus (1)
............ ........................................................................................... .. 180
Table 27.
Compound Psych Verbs which Derive with an Adjective (1) ............... 181
Table 28.
Compound Psych Verbs which Derive with an Adjective (2) ......... ;,: .... 181
Table 29.
Compound Psych Verbs which Derive with an Adjective (3) ............... 181
Table 30.
Compound Psych Verbs which Derive with an Adjective (4) ............... 182
Table 31.
Compound Psych Verbs which Derive with an Adjective (5) .. .. .. .. ...... . 182
Table 32.
A New Classification for Turkish Psych Verbs Including Compound Forms
......................................................................................................... 182
II. 5.
Table 1.
Psych Adjectives Derived with -Gan ............................. ... .......... 185
Table 2.
Psych Adjectives Derived with -GI + c/1 III siz .. ............................ 186
Table 3.
Psych Adjectives Derived with-gif ............................................ 186
Table 4.
Psych Adjectives Derived with -Ik ............................................. 186
Table 5.
Psych Adjectives Derived with --Inti+ (l/1 slz) .............................. 186
Table 6.
Psych Adjectives Derived with -Inr; + lll siz .................................. 187
Table 7.
Psych Adjectives Derived with-/~ + III slz .................................... 187
Table 8.
Psych Adjectives Derived with -It+ /II slz .................................... 187
XVI
Table 9.
Psych Adjectives Derived with -Ar + III slz ................................... 187
Table 10.
Psych Adjectives Derived with -eA +III slz .......... ......................... 187
Table 11.
Psych Adjectives Derived with -DA + II ....................................... 188
Table 12.
Psych Adjectives Derived with -sf ............................................. 188
Table 13.
Psych Adjectives Derived with -Ar doner ...................................... 188
Table 14.
Psych Adjectives Derived with -kaloz ......................................... 188
Table 15.
Psych Adjectives Derived with -DIRI -t + lei ................................. 189
Table 16.
Psych Adjectives Derived with- Ill n + mA + dlk ........................... 189
Table 17.
Psych Adjectives Derived with -An ............................................. 189
Table 18.
Psych Adjectives Derived with -II+ mA + mf$ .... ........................... 189
Table 19.
Psych Adjectives Derived with -DIRI -t + Jk + CI .......................... . 189
Table 20.
Psych Adjectives Derived with -Im + II+ sfz ................................ .190
Table 21.
Psych Adjectives Derived with -Ill n +An/ mAzl mAyan ................... 190
Table 22.
Psych Adjectives Derived with -1m/ n/ GI +sAl ............................. 190
Table 23.
Psych Adjectives Derived with -msAr ......................................... 191
Table 24.
Psych Adjectives Derived with -Ilk ......................................... .... 191
Table 25.
Psych Adjectives Derived with -C/ ........... .... :............................ 191
Table 26.
Psych Adjectives Derived with -AI+ t +lei .......................... ........ 191
Table 27.
Psych Adjectives Derived with -sAk ....... ................................... 191
Table 28.
Psych Adjectives Derived with -C/ ............................................. 192
Table 29.
Psych Adjectives Derived with -In ............................................. 192
Table 30.
Psych Adjectives Derived with -kar ................................... ......... 192
Table 31.
Psych Adjectives Derived with -$inas .. ....................................... 192
Table 32.
Psych Adjectives Derived with -i ............................................... 192
Table 33.
Psych Adjectives Derived with -sAl ............................................ 192
Table 34.
Psych Adjectives Derived with -Dan .......................................... 193
Table 35.
Psych Adjectives Derived with - !A + lei ................................... .. .193
Table 36.
Psych Adjectives Derived with -n{:
Table 37.
Psych Adjectives Derived with -II .............................................. 194
Table 38.
Psych Adjectives Derived with -slz ............................................ 194
Table 39.
Non- derived Psych Adjectives ................................................. 194
Table 40.
Psych Adjectives Derived with -lA$ ............................................ 195
............................................. 193
xvu
Table 41.
Psych Adjectives derived with -!An ............................................ 195
Table 42.
Psych Adjectives derived with -lA .............................................. 196
Table 43.
Lexical Counterparts of Complex Nucleus Verbs ( 1) ........................ 196
Table 44.
Lexical Counterparts of Complex Nucleus Verbs (2) ........................ 196
Summary Tables
Table 1.
Summary of the Causativity Properties of Turkish Psych Verbs ..... : ..... XIX
Table 2.
Summary of the Psych and Non-Psych Compound Forms ................ XXII
XVlll
THE LIST OF FIGURES
I. 1.
Figure 1.
Bresnan (1995) Lexicalist Views on Argument Structure .................... .22
Figure 2,
Bresnan (1995) The Argument Structure in LFG .............................. 22
I. 2.
Figure 1.
Belleti and Rizzi (1988) the temere class ....................................... 35
Figure 2.
Belleti and Rizzi ( 1988) the piacere and preoccupare classes .............. 35
Figure 3.
Belletti and Rizzi (1988) a Flat Tripartite Structure ........................... 38
Figure 4.
Belletti and Rizzi ( 1988) Theme Prominent Structure ........................ 3 8
Figure 5.
Belletti and Rizzi (1988) Experiencer Prominent Structure ................... 39
Figure 6.
Belletti and Rizzi (1988) Theme and Experiencer as a Small Clause ....... 39
Figure 7.
Kural (1996) Experiencer Subject Neutral Object Verbs ..................... 56
Figure 8.
Kural (1996) Neutral Subject Experiencer Object Verbs ..................... 56
II. 5.
Figure 1.
The Experiencer- Stimulus Interaction ....................................... 184
1
INTRODUCTION
Background Information
The relationship between the syntactic and the semantic components of
language has been one of the central issues and a matter of discussion in modem
linguistics. The validity and the coverage of the generalizations related to these two
components actually depend on a clear identification of the nature of the relationship
between the two.
In fact, the discussion point is not the question of whether there is a
relationship between the two or not but rather the extent of this interconnection. The
proposed answers, depending on the theoretical framework within which they are
presented, viewed the correlation between the syntactic and semantic realizations
sometimes as tendencies and sometimes as rules.
With a significant departure from his previous commitment about the
autonomy of syntax in GB, Chomsky focused on the relationship between the heads and
their syntactic complements which he called the thematic relations. The notion of thematic
roles was not something new at that time. It was first introduced by Fillmore and then
Gruber and recently by Jackendoffwho all have different claims at some crucial points.
Chomsky gives a more central position to thematic roles viewing their function
as the derivation of the subcategorization frames. This departure from his prior position as
to the autonomy of syntax was interpreted as an inevitable departure claiming that
" ... thematic roles are introduced into Chomsky's Government and Binding Theory only
2
because other syntactic conditions fail to block all the ill formed structures generated by
the grammar" (Ravin, 1990:12).
After his acceptance of thematic relations, thematic roles have become much
more popular and appeared in the theories which follow him (such as in the model of
Lexical Functional Grammar (LFG)). Under Chomsky's view, deriving syntactic structures
from semantics helped the linguistic theory in terms of simplicity and comprehensiveness
since it supplied the linguistic theory with no need to formulate two sets of principles; one
for syntactic transformations and the other for semantic combinations.
In the theory, reference to semantics is limited to Theta roles that are viewed as
the only semantic notions which bear syntactic relevance. To mean to be syntactically
relevant, a lexical item has to correspond to a syntactic behavior. Theta roles in the theory
have an abstract syntactic nature which does not have any semantic content and the
semantic information about lexical predicates is considered to be only in the form of
thematic roles in the lexicon.
The theory assumes that there are two properties in the determination of Theta
roles. The first is the properties that the lexical items have, and the second is the
grammatical functions which link the Theta role of the head with its argument. That means
it is possible to determine the syntactic configuration of a predicate from the thematic
roles. This view brings the idea that thematic roles also correspond to syntactic relations
between a verb and its complements. One of the current views is that the meaning of a verb
completely determines the configuration of the thematic roles. Based on this view, Theta
roles which express semantic relations between the predicates and their arguments were
regarded as semantic primitives.
3
A Projection Principle was also introduced in order to link the lexical items to
syntactic configurations. This principle assumes that once the lexical information is
projected onto the sentence, it can not be changed or deleted by any of the transformational
processes.
Together with the Theta Criterion and the Projection Principle, Chomsky
proposes a set of other principles which complement each other. The first one is the
Principle of Full Interpretation, when regarded from the point of view of predicate
argument structure, states basically that there should not be any superfluous elements i.e.
arguments of a predicate, in the interpretation of the sentences. This means the only
elements that can possibly exist in a sentence are the elements which have to be there. This
entails the Principle of Economy which requires the linguistic structures to be as
economical as possible.
One ofthecriticisms about the theory was.related to the syntactic and thematic
correspondence. Unlike the assumptions of the theory, a thematic role could exist in the
meaning of a verb without having a certain syntactic pattern as well as it was possible that
there may be some thematic roles which were absent in the meaning of a certain verb
though they were reflected in the syntactic structure of the sentence.
Another criticism was about the semantic contents of the thematic roles.
Although the Theta Theory claimed that Theta roles express the semantic relation between
predicates and their arguments, there is not anything about the semantic content of these
Theta roles themselves in the theory -since Chomsky uses Theta role labels just for
syntactic purposes.
The other criticism is about the concept of semantic class which does not have
any clear cut definition in the theory. Ravin observes that
4
.... The notion of semantic class is unclear- verbs have any number and kind of semantic
markers in common, thus forming a continuum of semantic similarity ranging from complete
dissimilarity to synonymy. It is not obvious how to break this continuum into distinct semantic
classes (1990: 226).
Actually the theory does not offer any criteria to determine the semantic
classes. Therefore, it is not obvious what criteria form semantic classes; whether
synonymy, near synonymy or some other defined senses which exist in the meaning of
some verbs will make lexical items share the same semantic class. In fact, the deliberate
omission of semantic discussions in the theory in general is one of the main points of
criticism as well as other criticisms such as about the criteria that make pragmatic
decisions different from the semantic decisions of native speakers to reflect their intuition
about grammaticality and acceptableness.
One of the other criticisms about the theory is that it bases itself mainly onto
considerations of language acquisition. The first aim of the theory was to represent the
universal principles and the nature of the linguistic competence. After Chomsky's shift of
attention from the autonomy of the syntax, to the lexicon and the lexical information,
during the course of the development of the theory, there has also been a shift of attention
in the area of language acquisition. Recently, much of the language acquisition has been
regarded as the acquisition of the features of lexical entries. As Chomsky himself points
out "language acquisition is in essence a matter of determining lexical idiosyncrasies"
(Cook, 1997:283).
In fact there are some other influential ideas concerning similar issues before
and after Chomsky. It is apparent that discussions after Chomsky are either in the form of
agreement or disagreement of his basic ideas.
5
One theory which accounts for the semantic relations between the predicates
and their arguments at a syntactic level is Fillmore's Case Theory. He defines six cases in
the theory; Agentive, Instrumental, Dative, Factitive, Locative and the Objective. These
cases, for him, represent just the concepts which are syntactically relevant, so,_~hey have
little to do with the semantics. In his view, the lexicon does not contain any semantic
information which inherently exists in the meaning of verbs.
There are also some criticisms about Fillmore's Case Theory. First, these cases
seem to represent just an arbitrarily chosen set of semantic properties, so, the theory
neglects some important meaning differences (such as the grouping inanimate objects and
natural forces under the instrumental case). Moreover, ·the cases listed do not cover many
semantic concepts inherent in the meanings of some predicates which causes him to ignore
some semantic nuances. Therefore, Fillmore's Case Theory was regarded as.an insufficient
modal classifying some arbitrary semantic concepts into some other arbitrary categories.
The model proposed by Jackendoff differs from both Chomsky's and
Fillmore's models in the underlying idea that it assumes two independent levels of
autonomous representation; syntactic and semantic. He gives the primacy to motion verbs
and sees them as the basis for the representation of all other predicates. He assigns five
· thematic roles in the analysis of the meaning of verbs as the; Theme, Source, Goal, Agent
and Location and analyzes the semantic functions of all other verbs based on motion verbs.
There are also some criticisms and counter ideas related to the model. First, it
is claimed that trying to force to capture the meaning of all predicates in terms of motion,
restricts the semantic coverage of the theory because there are a number of human
activities which do not involve the fundamental concepts of motion. Many lexical items
such as the psych predicates express a more abstract kind of motion which has a different
6
semantic nature. The theory is also claimed to misinterpret the semantic facts expressed by
the meaning of a number verbs. Secondly, it seems to be arbitrary to choose the motion
verbs and positing a GO function as primary, and trying to label all other verbs simply by
extending the notion of motion thus by giving the notion of motion some kind of a
privileged semantic status. The third criticism is that the theory has accounted for only five
thematic relations which are not convenient for the analysis of the meaning of some verbs.
To conclude, according to the criticisms, Jackendoffwas claimed to do wrong not only by
using insufficient thematic role labels to analyze the verbs but also forcing them into an
inconvenient motional framework.
As mentioned briefly, one of the criticisms for Chomsky's account related to
thematic roles was that the theory did not offer an elaborate semantic description of their
nature. The MIT Lexicon Project Theory (MLP) especially with Levin and
Ho~~v
shaped
the discussions about the Theta roles in the Government and Binding framework in a
slightly different way. The MLP linguists defend the idea that the semantic class that a
verb belongs to determines its thematic relations, in other words, the thematic structure of a
verb completely determines its syntactic behavior and the members of similar verb classes
show similar argument alternations. The correlations among semantic classes, syntactic
realizations of arguments, and the assignment of thematic roles are completely predictable.
They propose two levels of lexical representations; the first one is the Predicate Argument
Struc;ture {PAS) and the se.cond one is the Lexical Conceptual Structure (LCS). There are
linking rules which link these LCSs to the corresponding PASs. PASs are just like
subcategorization frames and LCSs are more semantic in their nature. The roles in PASs
are not marked as Agents, Patients etc. but they are structurally regarded as external or
internal. The Theta marking processes correspond to these arguments are the indirect and
7
the direct Theta marking. When the violations as to the linking of arguments of some
intransitive verbs came to be questioned, a distinction between two classes of intransitive
verbs which are unergatives and unaccusatives were proposed.
In the analysis of the extent of the interaction between the syntactic and
semantic components, the syntactic aspect has been a more investigated area in comparison
to the semantic component especially in the Generative Grammar Tradition.
The idea that the verb meaning is the main determinant in the syntactic structure
of sentences has caused the lexical semantics and especially the verb semantics to gain
more significance. It has been claimed that the semantic properties of a verb i.e. some
semantic concepts which are inherent in the meaning of verbs specify the syntactic
structure that the verb necessitates. The most remarkable reflection of the role of semantics
on the syntactic structure is observed in verb-argument relationships since the verb is the
main determinant of the semantic roles that the arguments in the argument structure can
bear. The verb determines both the number and the kind of arguments in a sentence. In this
way it is in the central position in both the constitution and the interpretation of sentences.
Kageyama (1997) mentions two main lexical approaches which are the
predicate-centered and the role-centered ones. Although the role-centered approaches
labeled the arguments with such roles as Agent, Patient, Goal and Location etc., they did
not say much about the meaning of the verb itself. The predicate centered approaches on
the other hand assume that verbs have lexical conceptual meanings. This approach isolates
the meaning of verbs from its cognitive aspects focusing on the structural properties.
The existence of some linking rules that link certain arguments which bear
certain semantic roles onto certain syntactic positions, and the idea that there are
regularities (linking regularities) that these rules have to obey, have led linguists to study
8
on possible linking theories. This idea has been claimed sometimes to mean that there is a
linking between the semantic and the syntactic components and sometimes to mean that
the two are completely interwoven.
!he. P~rpose of the Study
A group of psychological verbs present a kind of problematic area for the
mentioned linking or mapping theories according to the principle that thematically the
most prominent or the highest argument should also be syntactically in the most prominent
or the highest position. A significant number of the studies about psych verbs are about
these linking rules and the problems that a group of verbs of these verbs pose.
This unusual behavior of psych verbs were sometimes regarded as a matter of
difference in their deep structures and sometimes regarded as a matter of differenee in their
aspectual properties. Different approaches presented different explanations related to this
so called irregularity or exceptionality. However one of the important questions to be
answered is about the syntactic and semantic properties of this group of verbs in languages
other than the most thoroughly investigated ones in the literature. Therefore the aim ofthis
study is to describe the syntactic and semantic properties of these verbs in Turkish and
therefore contribute to the universal linking rules finding out cross linguistic
commonalities ifthere are any.
Apart from the universal linking rules, it has been expeCted thaftlie present
study may contribute to the description of Turkish grammar in some or other way with a
clear identification of some of the syntactic and semantic properties of psych verbs which
constitute a significant part of all verbs in Turkish that have received little or no attention
9
except for some of their semantic properties in traditional Turkish grammars. Therefore,
the study aims at contributing to the literature not only because it investigates a group of
verbs which are labeled in the literature as exceptional but also because it examines
Turkish whose psych verbs has not been studied so far.
Research Questions
The syntactic and semantic properties of psych verbs are gomg to be
investigated through the following questions:
1. Which verbs constitute the psych verb class in Turkish?
2. What kind of a verb classification can be proposed for Turkish psych verbs
when their syntactic and semantic properties have been described?
3. What structural and semantic. properties do the psych verbs in Turkish have in
terms of their interaction with voice markers?
4. What are the compound psych verbs in Turkish?
5. What are the structural and semantic derivational properties of the
adjectivalizations of psych verbs in Turkish?
Hypotheses
The following hypotheses are presupposed:
1. An analysis of the Dictionary of Turkish Language Institute will partially
provide the data needed to create a database of psych verbs in Turkish.
10
2. There is a psych verb classification with similar properties in Turkish with the
ones proposed in the literature for different languages when their syntactic and
semantic properties have been described.
3. The exceptional behavior of psych verbs are also observed in their interaction
with voice markers in Turkish parallel to the universal tendencies.
4. There are a number of compound psych verbs in Turkish which display similar
structural properties with their non- psych counterparts.
5. The derivation of the psych adjectives does not display any idiosyncratic
properties which differentiate them from general derivational patterns of
Turkish.
Data Collection Techniques and Limitations
The data for the psych verbs, psych compounds, psych nouns and psych
adjectives in Turkish is constituted of and limited to the verbs in the Dictionary of Turkish
Language Institution Volume . I and II (1988) and, the natural data needed for the
identification of subtle semantic differences is constituted of both the sentences from
electronic sources and of native speakers themselves.
The Method of Analysis
The data for psych verbs in Turkish is going to be analyzed according to their
syntactic and semantic properties creating a database to examine through a set of specified
criteria.
11
The methods of classification and description are going to be used with
comparisons and contrasts with the studies of the psych verbs in different languages in the
literature of which only the English translations will be given.
12
I.
I. 1.
REVIEW OF LITERATURE
Argument Structure
Why are there a limited number of arguments cross linguistically? What is the
restriction on the number of the arguments that a verb possibly can take? Why are there a
limited number of theta roles? This section analyzes the proposed possible answers that
constitute the fundamental concerns of argument structure theories.
The process of lexicalization is a language specific one in a way that it includes
historical phenomena as well as cultural ones. In spite of the differences in the lexicons of
different languages, there is the chance to express similar or sometimes even equal senses
by combining these lexical items with others and forming more complex structures in
different languages. Lexical items gain new meanings as a result of combinations of
different senses, moreover, new lexical items are often introduced into languages.
If the lexical items themselves expressed every aspect of events that they
denote, we would have as many lexical items as the number of all events in the world
which is not possible. This is also the case with the psych verbs. There are innumerable
psychological states of human beings each of which can be gradable in themselves. Thus
from the point of view of the nature of human computational system, it seems impossible
to denote each of these psychological states with a different lexical item. Rather we
combine some lexical items with others or we modify them to convey a number of
different senses.
Within the GB framework lexical structure has gained a more central role in
syntactic description. In this tradition, phrase structure rules became more redundant and
13
started to be eliminated. Movement was started to be assumed simply as the checking of
lexical features in syntax. Lexical Inclusiveness in the Minimalist Program proposed that
the output representations are just the properties of lexical items in the lexicon. The
derivation of sentences starts from these lexical items; which is the process called
numeration. These lexical items carry the syntactic, semantic, morphological and
phonological information. They merge together and create the constituent trees. When they
move to new positions, certain lexical features of them are checked according to the
checking requirements encoded in the lexicon.
Kageyama (1997) views verbs as the items which play the central role in the
composition and interpretation of the sentences since they determine the number, the kind
and the semantic properties of arguments that will exist in the sentence. In the analysis of
verbs, three levels of lexical representations are proposed up to now; the argument
structure, lexical conceptual structure and the event structure. Lexical semantic studies
have recently focused on two
approach~s
about the verb semantics and the syntax. They
are the role centered approaches and the predicate centered approaches. In the predicate
centered approaches the method of predicate decomposition is employed. Decompositions
in Kageyama's terms, are the representations showing only the skeletons of the verbs'
conceptual meanings (1997:5). These lexical conceptual representations are linked to
syntactic structures by means of linking rules.
After the recognition that there are significant regularities between argument
realizations of different predicates, the idea that the syntax of sentences is determined by
the meaning of predicates has gained popularity. These regularities are called ''linking
regularities" and the rules which map semantic roles onto syntactic positions are called
"linking rules" (Levin and Hovav, 1996:487). Some approaches assume that these
14
syntactic positions are unpredictable and some assume that they are partly predictable.
Levin and Hovav (1996) argue that this mapping between the semantic representation and
syntactic expression of the arguments of a predicate is fully predictable. They claim that
the best way to find out the syntactically relevant aspects of the meaning of a predicate is
to express the lexical semantic representation of predicates with a predicate centered
approach. To achieve this, we need to find out the elements of core meaning of predicates
which are essential in the linking of their arguments.
There are some cross linguistic differences in the mapping of arguments to
syntactic positions. According to Levin and Hovav, languages may even differ in the
linking of the arguments of two verbs which seem to be the translations of each other in
two different languages. Languages may also differ in the weight of the value of some
aspects of meaning that they give their arguments i.e. for the same argument, when one
language gives weight to one semantic component, another language may give weight to
another aspect of it.
Levin and Hovav mainly focused on the class-membership
r~l~tions
of
arguments. Under their view, if we group verbs into semantically coherent classes, we can
isolate the meaning components of a predicate relevant to syntax. Although some
idiosyncracy is allowed, generally, much of the syntactic behavior of verbs can be
determined by their membership in a semantic class. The basic suggestion is based on the
idea that the meanings of verbs have some kind of an internal structure. These internal
structures have some primitive elements. The similarities in the meanings of verbs are the
results of these shared elements in the decomposition of primitives.
Manning (1996) views the argument structure as a syntactic representation. For
him grammatical structure is a result of the grammaticization of discourse roles. For
15
example the notion of subject is the result of the grammaticization of the notion of topic or
focus of an event. That means there is an association between the topicality and
subjecthood. He states that we need two mappings between the gr-structure and a-structure.
The first mapping is the argument projection which is based on the meaning of predicates.
The second is the linking which links the argument structure to grammatical structure.
In recent years, event structure has also been seen as a modular component of
argument structure. It is even claimed to be responsible for the linking of arguments to
syntactic positions. Then, event structure was started to be represented in phrase structure.
They put the elements of event structure in VPs or in functional phrases. These ideas led to
the emergence of more syntactic analyses of the event structure.
Another property of event structure related to the argument structure is the
distinction between the individual level predicates and stage level predicates. Kratzer
argues that the stage level predicates have extra event arguments in their representations
(Tenny and Pustejovsky, 1990:20).
In the theories of verb meaning it is usually assumed that meanings of verbs
are multidimensional and the linking rules are sensitive to these dimensions. In the early
theories it was advocated that verbs that refer to the same kind of event with same
participants, display the same linking pattern. The verbs that do not conform to the
standard linking pattern were regarded as non canonical or marked. They were thought to
be rare in languages and difficult to acquire. Gropen and Pinker et al (1995) claim that
these so called non canonical verbs are in fact more numerous than canonical ones and
both are acquired at the same time.
16
Semantically, exammmg the argument structure of a verb is in a way
examining the entities and their relationships of these entities with each other specified in
the meaning of a verb. Ravin argues that
The arguments are place holders for entities. Since predicates have as many arguments as there
are entities represented in their meaning, whether these are syntactically realized or not, there
are four types of relations logically possible for arguments and complements .. . (1990:160).
He then lists these linking possibilities as; first, there are arguments that
inherently exist in the meaning of the verb but never have syntactic realizations, second,
there are arguments which are inherent and compulsorily syntactically realized, third, there
are arguments which are inherent but optionally realized, and the fourth there are some
other syntactic arguments which do not correspond to semantic arguments (1990:168).
According to Raves, the syntactic complements do not always correspond to
semantic arguments. On the contrary Chomsky has the claim that the grammaticality
determines the meaningfulness.
What Raves argues is that propositions can be both
grammatical and meaningless at the same time or vice versa. In fact, the idea that lies
behind all these arguments is the difference between the questions as to whether the
syntactic and semantic components are autonomous or they are dependent on each other.
The theories of argument structure mainly followed either of these approaches.
In Cross-Linguistic Patterns of Linking Valin (nd.) argues that the universal
aspects of argument linking exist and the language specific variations are rather limited to
a few number of possibilities. In the framework of Role and Reference Grammar, the
linking between the syntax and semantics is achieved at two levels. One is the lexical
phase level and the other is the syntactic phase level. Valin suggests that the universal
phases are found in the syntactic phases of linking but cross-linguistic variations
in the lexical phases of linking.
~refound
17
Wunderlich (nd.) argues that contrary to the assumptions, there are more
argument linking types than the generalized ones. He claims that neither of these types
belongs to UG but they are invented by cultural evolution. Some of these types are:
portmanteau type, active type, salience type, positional type and the inverse type. He sees
the inverse type as the most complex type since it maps two hierarchies on each other, such
as psych verbs.
The following section is a brief analysis of the the significant approaches to the
argument structure phenomena and the linking between the semantic arguments and the
syntactic realizations.
" ... what a theory of thematic roles should look like is analogous to that of the
blind men examining the elephant, each touching a different part of its body" (1991 :561)
says Dowty criticizing the traditional accounts of thematic roles and proposing an original
theory of his own called Thematic Proto Roles and Argument Selection.
What the earlier accounts on the same matter call; thematic relations (of
Gruber); deep structures and transformations (of Chomsky), subject sekctip!J _rules (of
Fillmore), template matching (of Stowell), universal alignment principle (of Perlmutter
and Postal), and universal theta assignment principle (of Baker), is called as Argument
Selectional Principles by Dowty.
Dowty claims that semantic distinctions are results of distinctions in the real
world. Thus, it is wrong to try to identify clear cut boundaries for these classes and to try to
find out the limits of our cognitive ability by refemng to those classes.
What he means by argument selection is not a kind of linking between the
syntactic andsemantic levels but a kind of constraint only on some lexical predicates out
of a great number of others. What he means by prototype is not about individual lexical
18
items but rather a generalization about lexical meanings. The argument selection principles
that he defines are only about two place predicates which have a subject and a true direct
object. Moreover, he takes only the 'arguments' into consideration but not 'adjuncts'.
Dowty sees thematic roles as prototypes or cluster concepts and argues against
viewing the thematic generalizations as equal with semantic and syntactic generalizations.
He defines five criteria (1990:572,573) for the properties of Agent Proto Role
which are volition, sentience/ perception, causing event or change of state of another
participant, movement and the independent existence of the event described by the verb. He
defines another five criteria for the prop·erties of a Proto Patient Role which are change of
state, incremental Theme, causally affectedness, relative stationariness and the dependent
existence of the event described by the verb.
The principles that he formulated as to the selections of arguments (1991:576)
can briefly be explained like this:
1. The argument which has the most proto Agent properties is lexicalized as the
subject.
2. The argument which has the most proto Patient properties is lexicalized as the
direct object.
3. If there are two arguments which have the same number of proto agent and proto
patient properties both of them can be lexicalized as subjects/objects.
4. If the predicate is a three place predicate, the argument which has the most proto
patient properties is lexicalized as the direct object, and, the one which has the less
proto patient properties is lexicalized as the oblique object or the prepositional
object.
19
5. lfthere are two arguments which have the same number of proto patient properties
both of them can be lexicalized as direct objects.
6. Some arguments may have none of these roles.
7. Some arguments may share the same role.
8. Some arguments may have the properties ofboth proto roles either in at?- equal or a
partial degree.
Dowty states that these principles are just strong tendencies rather than clear
cut rules. The combinations (1991 :571) for the corresponding roles that he foriritilated can
be summarized in the following table:
Ag~nt
....
';
Experiencer
lnstru~ent
Theme
Patient
volition+ causation+ sentience+ movement
volition+ causation
volition
sentience
causation+ movement
change+ (incremental Theme)+ dependent existence+ (instrumental Theme)
change+ (incremental Theme)+ dependent existence+ (cimsally affected)
Table L
Dowty (1991) Thematic Roles and the Syntactic Reahzatton Rules
For him, these roles and their argument selection principles determine the
following role hierarchies (Dowty, 1990:578):
Agent> Instrument/ Experiencer> Patient > Source/ Goal
Causing Event> Caused Event
Moving Argument> Source/ Goal Argument
On the other hand, Hale and Keyser (2002) define ·argument structure as a
lexical item's projection of its syntactic configuration which is limited both in variety and
complexity among languages. Similarly the number of thematic roles associated with them
is also very limited. They also claim that languages do not differ in their basic elements of
argument structure.
They define two basic structures as the complements and the specifiers
mentioning two argument structure types; lp-monadic (in which "lp" refers to the lexical
20
projection) which does not include a specifier but only a complement. On the other hand,
lp-dyadic predicates have a specifier argument. They point out that the subject argument is
not a specifier argument but rather an external argument. The specifier argument in their
view is internal in the lexical projection and shared by both the transitive and the
intransitive alternants since it is internal to the argument structure.
Alberti (1997) assumes that the internal structure of argument structures have
two substructures. The first one is the relative structure which is about perspective of the
speakers and the second one is the absolute structure which is the characteristics of a
whole family of related argument structures.
For him, the thematic roles are abstractions over semantic structures. The
important thing for an argument is not its thematic role but the position it occupies in the
internal structure of argument structure. Similar to Grimshaw (1990), he emphasizes the
importance of the ordering of arguments. The situation described by the verb is also
important. What he calls as the family of an argument structure is different argument
structure versions of the same verb. He claims that the internal structure of an argument
structure can easily be calculated on the basis of the family of the verb. This rule is valid
for canonical cases but there are also non canonical cases like psych verbs.
According to Alberti, the thing to do is to sort out the arguments which play the
central role in the situation described by the verb. Sometimes each participant in the
situation is not syntactically realized. He calls the overtly expressed arguments as explicit
arguments and non-overtly expressed arguments as implicit arguments. Other arguments
which are not the essential participants in the situation are called as adjuncts. Moreover,
some verbs may have more than one argument structures. Alberti assumes that the
argument version with a lesser number of arguments is prior to the version which has more
21
arguments. Therefore, speakers of a language have not only this knowledge but also have
the knowledge of different argument structure versions and the knowledge of the use of
one of these versions in different situations.
For Alberti, what Pesetsky's analysis makes clear is that in psychological
predicates either the Theme can be superior to the Experiencer or the Experiencer can be
superior to the Theme. Therefore, Alberti claims that it is not the thematic hierarchy that
we should look for but the hierarchy of arguments in the argument structure family. Alberti
adopts the argument hierarchy which was proposed by Larson and Baker:
Agent> Theme> Goal/ Benefactive/ Location
Baker On the Structural Positions of Themes and Goals points out that in some
languages Themes are more prominent than Goals and in some other languages Goals are
more prominent than Themes. Therefore it is not something easy to decide which one is
derived and which one is basic.
Bresnan (1995) in Lexicality and Argument Structure proposes a theory of
argument structure in LFG framework as opposed to other lexicalist frameworks such as of
Levin and Hovav and Hale and Keyser. Her main argument is that in a scheme including
the main components; lexical semantics, argument structure and syntactic structure, the
redundant element is not the argument structure · but the initial syntactic structure. That
means, in the linking process of the arguments to the final syntactic positions, the
underlying syntactic trees do not have any role and therefore they should be eliminated.
The following schemes illustrate the status of the mentioned initial syntactic structures in
other frameworks. The first one is of Hovav and Levin and the second one is of Hale and
Keyser (Bresnan, 1995:4, 5):
22
Lexical Semantics
!
Lexicon
Argument Structure
t
t
Syn~actic
Projection
Initial Syntactic Structure
Syntactic Transformations
Final Syntactic Structure
Lexical Primitives
t
Lexicon
Initial Syntactic Structure
!
Syntactic Transformations
Final Syntactic Structure
Figure I. Bresnan (1995) Lexicalist Views on Argument Structure
The last schema of Bresnan represents both her LFG framework and other
lexicalist views:
Lexical Semantics
!
Lexico Semantic Projection
Argument Structure
!
Lexico Syntactic Projection
Final Syntactic Structure
Figure 2. Bresnan (1995) The Argument Structure in LFG
She gives examples from Chichewa and English which support her view that
verbal processes for transitivity, intransitivity, causativity or passivization of verbs are
lexical morphological processes. Therefore, she claims that syntactic tree structures do not
have any functions in linking the arguments of verbs to their syntactic positions.
Another phenomenon which receives a great deal of discussion in the domain
of syntax semantics interface is the phenomenon of split intransitivity (unaccusative/
unergative distinction). The following section briefly analyzes some basic approaches to
unaccusative phenomenon.
23
I. 1. 1.
Transitivity and the Unaccusative Phenomenon
Unaccusative/ unergative distinction among languages have effects not only on
agreement and case marking but also especially on subject object distinctions and
therefore, has an important place in theories of argument structure.
According to the Unaccusative Hypothesis of Perlmutter (1978), some
intransitive verbs have two different classes as unaccusatives and unergatives. Unergative
verbs have an actual syntactic subject but unaccusative verbs have a derived subject which
is an underlying syntactic object.
Burzio states that there is a relationship between the case marking and the
argument structure (Hoekstra, 2000:55). There is also a correlation between the existence
of accusative case and the external argument. Unergative possesses external arguments but
not an internal accusative argument. Conversely, ergatives have internal arguments but not
external ones.
Alberti (1997) points out that in intransitive verbs the subject is either
associated with a patient like or an agent like grammatical function. The corresponding
language types are either ergative or accusative. He argues that there is a third type of
language (a mixed type) which assigns either an agentlike or a patientlike grammatical
function to the subjects of intransitive verbs. Accusative languages mark their intransitive
arguments with nominative case; just like the agent arguments of transitive verbs. Again in
accusative languages some intransitive subjects behave like agents and some intransitive
subjects behave like patients.
In some languages unaccusative and unergative verbs select different
auxiliaries, so, it is easy to distinguish them from each other. Case assignment properties of
24
these two groups of verbs are also different from each other. Burzio says that unaccusative
verbs do not have direct objects since they can not assign case for them but unergative
verbs assign accusative case to their _objects (Levin and _Hovav, 199():_492). .L.evtn and
Hovav also suggest that telicity as an aspectual property · does not always show
unaccusativity since there are verbs which are both atelic and unaccusative.
According to Chomsky, there is a differentiation between AGRs and AGRo.
The choice between the two is a matter of activation. Nominative- accusative languages
activate AGRs but ergative- absolutive languages activate AGRo. AGRs is seen as related
to Tense while AGRo is seen as related to V. Tense case is a property of all sentences but
verb case is not; that is why accusative case is not always available for all verbs. This is
claimed to be just a matter of asymmetry between the nominative and the accusative
(Hoekstra, 2000:58).
As mentioned, there are also different conceptions as to what the argument
structure is and at which level it should be analyzed. Manning ( 1996) argues for a
distinction between the grammatical relations and the argument structure prominence. In
Government and Binding literature, Grimshaw (1990) claims that argument structures are
not sets or lists of arguments of a predicate. They have their own internal organizations,
they affect the grammatical behavior of predicates and they are predictable from the key
characteristics of the meaning of predicates. Some sort of prominence relations ·a.mong the
arguments are determined by the thematic and aspectual properties of each predicate.
Manning ( 1996) suggests six classes of treatments of ergativity which are the
syntactic accusativity analysis, ergative as passive analysis, absolutive subject as object
analysis, the oblique analysis, the inverse analysis and the four relations analysis (37).
25
Manning adopts the inverse analysis and assumes that there is an accusative
organization for all languages at the level of argument structure.
He argues for the existence of at least three linking possibilities cross
linguistically. What he calls syntactic accusativity displays the following mapping
possibilities:
There are some languages which always use inverse mapping for transitive
verbs. These languages are what he calls syntactically ergative:
Gr- Structure
A- Structure
Object
A-subject (agent)
Subject
Patient
Table 2.
..
Syntactic Ergativity
He sees the argument structure as a syntactic level and valence changing
operations as operations on this level.
In brief, the basic claim of Manning is that the syntactic representation has two
levels of information which are the grammatical structure and the argument structure.
According to him, there are two different mapping possibilities between these two levels.
Manning introduces an Inverse Grammatical Relations Analysis and claims that these two
levels pose different sort of constraints on syntax and linking of these two levels differ
from one language to another.
Dowty (1991) sees the unaccusative/ unergative distinction as a grammatical
distinction and deals with syntactic accusativity and semantic accusativity separately. In
his analysis of thematic proto roles, Dowty predicted that his argument selection principles
do not apply to syntactically ergative languages since they have an inverse correlation
between the subject and object and thus between the proto agent and the proto patient.
26
However he claims that these principles can be applied in the same way just by means of a
reversal. In ergative languages, the arguments with most proto patient properties are
lexicalized as subjects and the arguments with the most proto agent properties are as
objects. His observation is that the unaccusative verbs have arguments with ·patientlike
meanings while unergative ones have arguments with agentlike meanings.
He argues that the most important proto property to distinguish between the
unaccusatives and unergatives is volition. In his analysis, the verbs which have the
arguments with the properties of sentience and volition are unergative verbs.
Examining "variable behavior verbs" which are the verbs that sometimes show
unaccusative and sometimes ergative behavior, Levin and Hovav (1996) claim that
unaccusative and unergative classes can be predicted semantically. Moreover they observe
that these verbs most of the time have more than one meaning i.e. when they display
unaccusative behavior, they have a different meaning and when they are unergative, they
are associated with another meaning (Levin and Hovav, 1996:490, 491).
Chung (nd.) proposes a number of criteria to distinguish between the
unaccusative and the unergative verbs. First, in terms of argument structure, unergative
verbs take external arguments while unaccusative verbs do not. Second, he claims that in
the events described by unaccusative verbs, the notion of intention is not involved; on the
other hand, unergative verbs describe a kind of willed intentionality. Third, the nominal
suffix -er is attached only to verbs which can take external arguments i.e. unergatives.
Fourth, the resultative construction can only be used with unaccusative verbs but not with
unergatives (12).
27
Nakipoglu (1998) argues that Turkish is one of the languages which are
sensitive to the unaccusative/ unerg<l.tive distinction. She defines three classes of
unaccusatives in Turkish which are the endpoint, measure and path unaccusatives.
She observes that semantic notions such as agentivity, volition, control and
delimitedness are important for the unaccusative/ unergative distinction for Turkish
intransitives too. Nakipoglu uses adjectival passives as a diagnostic to distinguish between
unaccusatives 'iud unergatives in Turkish. For example, -ml~ is incompatible with
transitives and it modifies only the subjects of unaccusatives. Another adjective deriving
suffix -Ik is also only compatible with unaccusative verbs. What Nakipoglu (1988)
observes for Turkish unaccusatives and unergatives can be summarized in the following
table:
Unaccusatives
Unergatives
externally instigated
can not be passivized
delimited
the only argument is affected and changed
internally instigated
can be passivized
non-delimited
the only argument is affecting or instigating rather
than affected
lack a direct internal argument
describe activities
can never select an external argument
describe situations
Table 3.
Nak1po~lu
(1998) Unaccusattves and Unergattves
Another grammatical process which constitutes a significant part of the
discussions in the psych verb literature is causativity. The following section briefly
summarizes some of the basic discussions about causativity.
I. I. 2.
Voice Alternations and Causativity
The idea that the causative meaning has an affect on the realization of the
arguments of a predicate supported the claim that the unusual realization of the arguments
of the Experiencer object psych predicates is because of their causative semantics.
28
Therefore, psych verbs which constitute a borderline area for lexical semantic and
syntactic studies have been analyzed mostly with reference to causativity.
Palmer (1994) sees causative constructions as derived from simple non
causative sentences just like passive constructions which are derived from their active
counterparts. Unlike passive constructions, causative ones add a subject Causer argument
to the argument structure. Just like many languages which have morphological passive
forms, some languages have morphological causative fonns like Turkish. Other than
morphological forms there are periphrastic forms which utilize a specific helping verb for
causation. For example English does not have any grarnmaticalized causative morpheme
but it rather uses periphrastic verbs.
There are four principles for causative constructions; first, there sh~uld be a
morphological or periphrastic mark on the verb, second, there should be a Causer addition
to the subject position, third, other arguments should be demoted, and fourth, there should
of course be a causative meaning. For "demotion" Palmer adopts Comrie's (1976)
hierarchy of grammatical relations (Palmer, 1994):
Subject> Direct Object> Indirect Object> Oblique Constituent
Palmer claims that the accusative case expresses direct causation which he calls
coercive causation but the instrumental expresses indirect causation which he calls
noncoercive causation. For him, dative case ts used with verbs of experience whose
subjects are typically animate non Agents.
29
He defines two types of causation which are the manipulative causation
(lexical/ single event causation) and directive causation. Manipulative causation
necessitates the physical movement of the Causer while directive causation does not.
In the causative constructions in Turkish, as mentioned, there is a demotion in
the hierarchy of grammatical relations in which the subject (SU) is the highest, direct
object (DO) is the second highest, the indirect object (IO) is the third highest and the
oblique object is the least high. Zimmer (1988) argues that in Turkish causative
constructions of intransitive verbs, the subject slot is filled and the Causee argument is
assigned the next highest case appropriate (which is the case of the DO). This is what
Zimmer refers to as "Syntactic Demotion Strategy (SDS)" (1988:217). Turkish is one of
the languages that uses this strategy. Zimmer claims that in Turkish this rule does not apply
regularly.
Turkish transitive verbs take accusative objects. With a small class of verbs
which take dative case marked objects, the mentioned rule does not apply. Zimmer calls
them as "case switching causatives" (219) exemplifying a similar process in French and
emphasizing that case switching is possible in the process of causativization only for this
small class of verbs. Therefore, we can neither see it with the verbs which have accusative
objects nor change or switch the cases ofNPs for topicalization purposes.
The discoursal effect of this process is similar to the passivization which
topicalizes the direct object. In these sentences too, the surface DO is more topical,
affected or patientlike element than the surface indirect object. Thus, Zimmer claims that
Turkish, with these possibilities in case assignment, allows its speakers to convey some
differences in the way they regard the roles of the elements in the sentence.
30
In brief, Zimmer's basic underlying claim is that there is a semantic flexibility
in the assignment of cases according to the communicative needs of the speakers.
Zimmer's analysis of case switching causatives is in a sentential level and there
is also a need for an analysis of the semantic concepts inherent in the meaning of such
verbs. A detailed analysis of the relationships between the semantic and syntactic
properties of these verbs would shed a light on the issue from a lexical point ofvlew.
Levin and Hovav (1996) divide verbs into two groups in terms of their
causativity as internally caused eventualities and externally caused eventualities. They
define a linking rule which relates Causers with the deep structure subject position which
they call a Causer Linking Rule (1996:501). For them, Agents and Causers are not the
same arguments. However, since an Agent is also responsible for the eventuality of the
verb, it is also a kind of Causer argument.
In decompositional approaches of verb meaning, Cause is a dyadic predicate
which needs a Causer and a Causing Event. According to Levin and Hovav, the Causer
argument may be an argument of an intransitive verb or a stative verb. Causer argument
can also be an Agent as well as a natural force. Moreover, they identify unergative verbs as
externally caused intninsitives and unaccusatives as internally caused intransitives.
Levin and Hovav define another rule; the Theme Linking Rule which links the
Theme argument to the deep structure argument position and argue that Theme Linking
Rule has precedence over Causer Linking Rule (1996:502).
Holmes's (nd.) analysis is a Word Grammar (WG) analysis of the syntax and
semantics of causative/inchoative alternation. In the WG analysis it is claimed that " ...
both causative and inchoative uses must be associated with independent lexical
31
representations since there are cases where both sides of the alternation have idiosyncratic
semantic properties" (Holmes, nd.:327).
The idea that there are events having complex internal struc;tures divided the
Theme argument into two parts as the inner event and the outer event. According to Tenny
and Pustejovsky the outer event relates to causation and agency, however, the inner event
relates to telicity and the change of state (2000:7). Simply, the outer event causes the inner
event. There are different linguistic approaches as to what causation represents. According
to an approach, it represents the relationship between the two propositions. On the other
hand, some approaches claim that it is the relationship between the two events and some
others claim that it is between an Agent and an event. For example, Jackendoff sees it as a
relation between an individual and an event or between a thing and an event. Dowty on the
other hand, sees it as a relation between two propositions. In Grimshaw (1990) we see the
activity as the outer causing event and the state as the inner event.
Another issue in question in the literature is the relationship between the Cause
and the Agency. Both Causers and Agents are syntactically realized as subjects. However
both can occur independently from each other. That means although these two notions
often intersect, they should be represented separately. One other argument about agentivity
is that most of the time Agents are regarded as volitional unlike Causers.
In the psych verb literature first we see Belletti and Rizzi (1988) ·who identify
three classes of psych predicates in Italian. According to their classifications Class I verbs
have nominative Experiencer subjects and accusative Theme objects (as in John fears
bears, Class 2 verbs have nominative Theme subjects and accusative Experiencer objects
(as in The bears frightened John), and Class 3 verbs have nominative Theme objects and
dative Experiencer subjects (as in Bears appeal to John).
32
The mappings of the arguments of these three classes are; Class 1 verbs link the
Experiencer role to the external argument. Class 2 verbs link the external argument to the
object position licensing a Causer role with causative morphology as opposed to the Class
1 verbs which are non causative and stative. Nelson (nd.) argues that the difference
between the two classes is due to the differences in causation and aspect as Grimshaw
(1990) argues. The Class 3 psych verbs differ from the other two classes. They are similar
to Class 1 verbs because they are stative too. They are similar to the Class 2 verbs because
they assign nominative case to Subject and object case to Theme but they are not causative
unlike Class 2 verbs.
Grimshaw (1990) argued that in causative psych predicates, the thematic role of
the subject is not the Theme but the Causer and the Causers are ranked more highly than
the Experiencers in the thematic hierarchy.
Fabienne (nd.) in Object Experiencer Psych Verbs (OEPVs) and Causatives
with supporting data from French, argues that not all object Experiencer psych verbs are
causatives. As known, causative verbs describe two events; namely, a causing event and a
caused event. If all OEPVs are causatives they should describe a causing and a caused
event but this is not the case for French OEPV s. The caused event is the mental event
affecting the Experiencer. This event can be either dynamic or stative. Fabienne claims that
some French OEPVs lack a causing event saying that only one class of OEPV s has two
eventualities similar to causatives.
Pylkkanen (2000) in Stativity and Causation argues against the idea that
stativity and causation are incompatible notions with evidence from a class of Finnish
psych predicates which are stative both in their causative and non causative uses. Her first
argument is that the Experiencer object predicates in Finnish are morphologically causative
33
but aspectually stative, just like non causative Experiencer subject predicates. Her second
argument is that morphologically non causative psych verbs refer to simple individual
states (i-/evel) but morphologically causative psych verbs refer to complex stage level
states (s-Ieve!). She demonstrates that causativity is compatible with stage level stativity
but not with individual level stativity in Finnish. Pylkkanen assumes that this type of psych
verbs are not derived from non causative psych verbs but rather they have their own
external arguments. She adopts Pesetsky's approach that the subjects of frighten verbs
have the Causer role but objects of fear verbs have the Target role and the Causer role is
received from the causative morpheme not from the predicate itself.
I. 2.
Approaches to Psych Verbs
The domain of emotions is such a rich field that it may even include a number
of other disciplines like linguistics, psychology, sociology and even anthropology.
Apart from its psychological and cultural aspects, a detailed linguistic analysis
of psychological verbs and cross linguistic evidence from a variety of languages will
provide a better understanding of the linguistic realizations of emotions and the nature of
the lexicon.
This section analyzes some of the significant approaches to psych verbs in the
psych verb literature. These approaches can be grouped under two general headings;
structural and semantic/ aspectual.
34
I. 2. 1.
Structural Approaches to Psych Verbs
Among the structural approaches, Belletti and Rizzi ( 1988) is the most
significant and mostly referred study which explains the exceptional behavior of a group of
psych verbs from a syntactic point of view. Pesetsky (1987) focuses on the binding
peculiarities observed with a group of psych verbs. Vanhoe (2002) analyzes Spanish psych
verbs in LFG framework. Mulder (1992) analyzes ergative and unergative psych verbs in
Dutch. Hale and Keyser examine the transitivity alternations of psych verbs in English.
Ginnis analyzes the morphological restrictions on causative psych verbs in English and
Kural (1996) analyzes Turkish psych verbs through elementary predicates. The following
sections briefly summarize the mentioned structural approaches to psych verbs.
I. 2. 1. 1. Belletti and Rizzi (1998) and the Psych Verbs in Italian
Belletti and Rizzi (1988) analyze the so called problematic area of psych verbs
which seem to constitute a resistance to the regular mapping of thematic roles to dstructures and identify three types of psych verbs in Italian which present different
syntactic configurations:
(1) Gianni fears this.
(2) This worries Gianni.
(3) a. To Gianni pleases this.
b. This pleases to Gianni.
They call these three classes as the Iemere, preoccupare and ·the pia cere
classes. The piacere class is different from the other two classes. It has a dative
35
Experiencer and a nominative Theme. This class allows both orderings of these two
arguments unlike others.
Traditional studies view the temere class as the main structure from which
other two structures were derived transf()rmationally.
Belle~ti
and Rizzi p;r_opo&~JP.a~ th~
4-
structure configurations of these three classes are not so much different from each other.
The first tree diagram is ofthe example (1) and the second is of(2) and (3):
s
~
NP
VP
A
V
NP
I
Gianni
teme
questo
Figure I . Belleti and Rizzi (1988) the Iemere class
s
v·
V
e preoccupa
piace
NP
NP
questo
Gianni
a Gianni
Figure 2. Belleti and Rizzi (1988) the piacere and preoccupare classes
The commonalities between these two diagrams are:
•
Both verbs directly theta mark the Theme argument
•
Verb+ Theme theta marks the Experiencer.
36
Fitsr they claim that the subject of the preoccupare class is riot adeep subject
but a derived subject. In order to prove this they use the anaphoric cliticization test. The
assumption is that the deep subject can bind a reflexive clitic while the derived subjects
can not. While the verbs in temere class are perfectly grammatical verbs of preoccupare
class are not:
(4) Gianni himself fears.
(5) *Gianni himself worries.
With the addition of a human subject, a reciprocal and an adverb like
intentionally or voluntarily, the structure becomes acceptable with a meaning that a human
subject is intentionally causes a psychological state in the Experiencer:
(6) These two guys frighten each other intentionally every time that they have
ADV
the opportunity.
The conclusion is that if the Theme of a psychological verb is non-agentive
then their subjects are derived.
Another test to prove the same claim is the causative construction test. Basing
themselves on Burzio's ideas, Belletti a_nd Rizzi claim that in Italian, whenever ..there is a
derived subject, the structure can not be embedded under a causative construction. While it
is possible for temere verbs to have acceptable counterparts embedded with a causative
verb (7), it is not the case with the preoccupare class (8):
(7) This caused that Mario him feared even more.
(8) *This caused that Mario him worried even more.
Moreover, causative verbs can also have their VP complements. The
observation is that temere verbs are possible with the infinitival verbs however, not all but
most of the verbs of preoccupare class are not:
37
(9)
This will make (one) fear the president even more.
(10)
*This will make (one) attract the president even more.
Belletti and Rizzi observe that this behavior of preoccupare verbs with
infinitival VPs is similar to unaccusative verbs.
Turning to their passivization behavior, Belletti and Rizzi claim that verbs of
preoccupare class do not have syntactic passives. The possible passive sentences with
these verbs are instances of adjectival passivization:
(11)
Gianni is disgusted by the corruption of this country.
(12)
Gianni is fascinated by this perspective.
because they:
•
show typical adjectival morphology
•
can not bear clitics (since only the verbs and therefore verbal passives can
bear clitics)
•
select the auxiliary essere 'be' (which is compatible with both adjectival
and verbal passives) but not venire 'come' (which is compatible only with
verbal passives)
•
do not allow regular participial form:
(13)
•
*I am tired/ excited by his ideas.
have irregular adjectival forms:
(14)
I am tired/ excitedofhis ideas.
Belletti and Rizzi see this as a result of a blocking principle which predicts that
the existence of an irregular adjectival form blocks the regular adjectival participle
formation (1988:313).
38
Another special characteristic of the verbs of preoccupare class is with
respect to their behavior in the theory of Binding; the Experiencer argument of these verbs
violates the c-command requirement for antecedent-anaphor relationship, binding an
anaphor contained in the subject:
(27)
*The gossips about himself describe Gianni better than any official
biography.
(28)
The gossips about himself worry Gianni more than anything else.
While the object of psych verbs can bind an anaphor which is contained in the
subject they can not bind an anaphor which is the subject itself:
(29)
Pictures ofhimselfworry him.
(30)
*Himself worries him.
Belletti and Rizzi claim that the subject of these verbs moves to the subject
position from an internal VP position. They propose four different possible internal
structures for this VP (1988:320):
1.
A "flat triparite" structure:
VP
V
Theme
Experiencer
Figure 3. Belletti and Rizzi ( 1988) a Flat Tripartite Structure
This can not be the right structure because the branching is not "binary".
2.
The Theme is more prominent than the Experiencer:
VP
V'
A
V
Figure 4.
Theme
Experiencer
Belletti and Rizzi ( 1988) Theme Prominent Structure
39
This is not the right structure because the Experiencer can not c-command the
Theme.
3.
The Experiencer is more prominent than the Theme:
VP
V'
Experiencer
~
V
Figure 5.
Theme
Belletti and Rizzi (1988) Experiencer Prominent Structure
This is the only structure that is compatible with the evidence of Belletti and
Rizzi.
4.
The Theme and the Experiencer form a small clause:
VP
Theme
Experiencer
Figure 6. Belletti and Rizzi (1988) Theme and Experiencer as a Smalt Clause
This can not be the right structure because there is no theta role for a small
clause constituent.
The Experiencer argument ofpreoccupare class is in the object position but it:
•
does not have the properties of canonical objects.
•
is a kind of secondary object.
•
is the sister ofV'.
•
is immediately dominated by VP.
•
is not transparent to extraction unlike canonical objects:
40
•
has some object properties (it receives accusative case)
Belletti and Rizzi claim that this accusative case is not the ordinary accusative
case which we see in the object arguments of simple transitive verbs. The accusative case
for them has two realizations which are structural and inherent. The general rule is this:
V is a structural case assigner iff it has an external argument.
The verbs of preoccupare class are not structural case assigners since they do
not have external arguments. Therefore this accusative case is an inherent case. They
observe that the auxiliary selection is also affected by this:
•
If a verb takes avere, it has an external argument. If it has an external
argument, it is a structural case assigner and has a deep subject.
•
If a verb takes essere, it does not have an external argument. If it does
not have an external argument, it is an inherent case assigner and it has a
derived subject.
The third class is the piacere class. Piacere class has the following properties:
•
Their Experiencer is in the inherent dative case.
•
They select the auxiliary essere.
•
Theme-Verb-Experiencer and Experiencer-Verb-Theme are possible
orders.
•
The most natural word order is the second one which seems to be
unmarked.
•
They are ergative verbs.
•
Their subject position is non-thematic.
•
Their Theme and the Experiencer are VP internal.
41
•
Their Experiencer is higher than the Theme. (higher is to mean ccommanding)
•
Either their Themes or their Experiencers should be in preverbal
position. Both should not remain in the VP.
•
Verb-Theme-Experiencer order is ungrammatical (as it is the case with
the temere and preoccupare classes) which is fully grammatical with
non-psych verbs:
(31)
*Please/ worry/ fear your ideas to Gianni.
(32)
Won the elections a student.
As a result Belletti and Rizzi draw the following conclusions from the analysis
of three types of psych verbs in Italian:
•
These three classes are identical to each other except for the selection of
different inherent cases.
•
The Experiencer is projected to a higher position than the Theme
(1988:344).
•
The lexical representations of three classes of psych verbs are (Belletti and
Rizzi, 1988:344):
<Class
temere
preoccupare
piacere
•
All s-structures which have an inherent case marked NP in their subject
positions have a derived subject
•
If a theta grid has both an Agent and an Experiencer, the external role is
always given to the Agent.
42
•
If there is no other highly ranked role than the Theme, or if the Theme is the
only member in the theta grid, it can be chosen as the external argument.
Otherwise a Theme can never be the external argument.
•
According to these generalizations, the following verb classes would not be
possible which:
a.
are like temere class but with no external argument.
b.
are like preoccupare class but with a simple transitive structure.
c.
are like piacere class but with structural accusative case assigned to
d.
are like piacere class but with avere as an auxiliary.
Theme.
•
According to these generalizations, the following verb classes would be
possible which:
a.
assign inherent case for both the Experiencer and the Theme
b.
assign inherent case for the Theme and externalize the Experiencer
I. 2. 1. 2. Pesetsky (1987) and the Binding Problems with Experiencer
Verbs
Am_o ng other peculiarities t~at Experiencer verbs pose, Peset~Jcy ( 1.9_87) deals
with anaphoric binding problems such as the following:
(1) Pictures ofhimselfl know John likes e.
(2) Pictures of each other annoy politicians.
(3) Each other's health worried the students.
43
The problem is that the anaphors in the sentences (1 ), (2) and (3) are not ccommanded by the antecedents in italics. It is important to note that the violations like
these are only seen with the verbs which assign the Experiencer theta role to its object.
Pesetsky observes that there is one more unusual property of Experiencer
Object psych verbs; these verbs license an infinitival clause with Tough Movement (TM).
In these constructions PRO in the infinitival clause is controlled by the Experiencer
argument of the matrix verb as in (4) and (5):
(4) These pictures(i) annoy mem [ PRO(i) to have to look e(i)].
(5) War(i) frightens meu> [PROm to think about e(i)].
but with non-Experiencer predicates it is ungrammatical:
(6) *Bill(i) kicked mem [PROm to have to look at e(i)].
Pesetsky observes another peculiarity related to the meaning of sentences with
Experiencer arguments. In order to interpret a sentence like John kissed Mary, we just need
to know if John kissed Mary or not. We do not need to know other circumstances -aboutthe
mentioned event. However, in order to interpret such a sentence like The article angered
John we need to know what exactly is the thing that angered him. He calls this as
"expressively incompleteness" (1987:130). Moving on from this point, Pesetsky predicts
that there are deleted infinitival clauses in these sentences which have undergone TM. For
him, the clause deletion is due to the discoursal factors. Therefore, the sentences in (7) and
(9) are clause deleted counterparts of (8) and (1 0):
(7)
Pictures of each other annoy the politicians.
(8)
[Pictures of each other (i) ]G) annoy the politicians(i) fPRO(ifto look at
[ecD]].
(9)
Stories about herself generally pleased Mary.
44
(10)
[Stories about herse(fii>] generally pleased Mary(i) [PRO(i) to hear [eJ
(j) ].
The anaphor inside the subjects is bound by the Experiencer object which is
non c-commanding. Pesetsky's assumptions related to this phenomenon are:
•
Experiencer Object psych predicates take an infinitive clause which
undergoes TM.
•
This infinitive clause can be omitted.
•
PRO in the tough infinitive clause is controlled by the Experiencer.
•
e in the tough infinitive clause is c-commanded by PRO.
•
TM causes some connectivity effects.
•
As a result of this connectivity effect, the subject NP which includes the
anaphor acts as if it is in the position of e.
•
The immediate antecedent of the anaphor is the PRO.
Pesetsky claims that Experiencer predicates have always Agent-Patient
counterparts and that these Agent-Patient usages exclude TM. Thus, in non Experiencer
sentences we neither see TMs nor peculiar binding effects:
(11)
Bill deliberately annoyed me (*to talk to).
(12)
The actor deliberately frightens the children (*to look at).
(13)
*Each other's friends deliberately annoyed/ frightened the party-
goers by blowing smoke in their faces.
Pesetsky's claim that Experiencer psych predicates contain a deleted infinitival
TM is only about the Experiencer predicates whose Experiencers surface as s-structure
objects and does not include all Experiencer predicates.
45
Another claim
IS
that if the sentence involves TM, the nominalization
IS
ungrammatical:
(14)
*the book's annoyance of John.
(15)
*the book's amusement of children.
Nominalization is possible with the Agent-Patient versions of these verbs when
they do not contain an infinitival clause:
(16)
Mary's (deliberate) annoyance ofJohn.
(17)
Mary's (deliberate) amusement of children.
Another observation is that the tough infinitive clause excludes reflexives as in:
(18)
?I disgusted myself.
(19)
?l(i) disgusted myself(i) [PRO to think about e(i)].
Pesetsky identifies two approaches as to the question of how many arguments
these Experiencer verbs can have. According to the first approach, there are three
arguments which are; an external NP, an Experiencer and an infinitive clause. The second
approach does not regard the surface subject as an argument of the verb and identifies two
arguments which are an Experiencer and an infinitive clause. Pesetsky notes that there are
some verbs like amuse which require the existence of an infinitive clause "independent of
TM":
(20)
It amuses Mary to read this book.
(21)
This book amuses Mary to read e.
(22)
*This book amuses Mary to read War and Peace.
These are similar to raising structures. The subject position may also be
occupied with an expletive element:
(23)
It seems that book has been read.
46
(24)
That book seems to have been read.
(25)
*That book seems that War and Peace has been read.
When
passivization
is
considered,
Pesetsky
observes
that
some
nominalizations seem better than others with passives:
(26)
??Bill was delighted by Fred.
(27)
Bill was delighted by Fred's visit.
(28)
??Mary was amazed by my shoes.
(29)
Mary was amazed by my choice of shoes.
According to Pesetsky, (27) and (29) can be derived without TM. The reason of
(26) and (28)'s seeming odd is because of the Causer argument. This argument can either
refer to an action or an event. But here, a shoe is neither an event nor an action therefore, it
can not be a theta marked argument of the verb but the choice of shoe can be.
Pesetsky notes some thematic and selectional differences betwe-en verbal,
adjectival and nominal examples. Therefore, there should be a distinction between the
Cause of Emotion and the Target of Emotion:
(30)
Bill was angry at the article in the Times.
(31)
The article in the Times angered Bill.
The sentence (31) differs from (30) in that the anger of Bill is not directed at the
article but something in the article. However, when we use the adjectival form (30), we
mean that the article itself is the object of his anger. Pesetsky claims that such pairs like
anger-angry, fear-frighten are not simple lexical variants.
The non Experiencer NP is sometimes the Object of Emotion and sometimes
the Cause although it is possible to think of a Cause also as the Object of Emotion as in
(32):
47
(32)
Bill's behavior embarrassed Sue.
Pesetsky's prediction is that " ... the noun and the adjective do not assign the
Cause role because the related verb is a lexical causative ... " (1987:137). He points out
that the causative verb can not assign the role of the Object of Emotion as, the
ungrammaticality of (33) shows:
(33)
*The article in the Times angered Bill at the government.
To conclude, Pesetsky's main argument is that it is possible to explain the
peculiar binding properties of Experiencer object psych verbs with connectivity effects and
with the existence of TM in these verbs.
I. 2. 1. 3. Vanhoe (2002) and the Psych Verbs in Spanish
Vanhoe (2002) analyzes the syntactic properties of psychological verbs m
Spanish in the framework of LFG. He distinguishes three classes of psych verbs m
Spanish. Each of these types presents a different correspondence pattern according to their
thematic roles and syntactic positions. The first group is the temer class which is similar to
the fear class in English. The second group is asustar or preoccupar class which is similar
to frighten class in English and the third class is gus tar class with a dative Experiencer.
In three of these classes, there is an Experiencer who reacts to a Theme
emotionally. Vanhoe argues that aspectually, the temer and the gustar classes are states but
preoccupar 'Class is achievement. States also differ in themselves; temer class behaves like
ordinary transitive verbs but gustar class shows unaccusative behavior.
He modifies Dowty's proto -Agent and proto Patient properties and analyzes
three groups of psych verbs in Spanish according to them:
48
V anhoe discusses another phenomenon called "leismo" which means that a
direct object may sometimes be marked with DAT. Most of the time it is optional,
pragmatic or contextual in Spanish. However there are also some verbs in Spanish whose
direct objects are obligatorily marked with DAT such as interesar. evidence
I. 2. 1. 4. Mulder (1992) and the Unergative and Ergative Psych Verbs
According to Belletti and Rizzi's analysis, the well known three types of psych
verbs have similar underlying structural representations. Mulder (1992) states that these
three sentences have radically different deep structures. The only similarity between them
is that they all have an Experiencer argument. Other than this feature, .th~se y~rb~ have
very little things in common.
He refers to the second type (frighten type) of verbs as s-psych verbs and the
third group as o-psych verbs which have the following properties:
- 8-;' P,SY~H (]ONSTRUCJJQNS~'i'>;
. . . ·E:&i)i.tit,z,: ~Q-l(:f,§,¥ClJ~"Q~~IRQg,!J~:n~.~~·~.'·t}Y'MAJ::£u·ii&'t':. -~ .
The Experiencer is the subject of the state of mind.
The Experiencer (the dative NP) is the subject of an
inalienable possession construction and it functions
as the possessor with a covert possessee.
The Theme is the causal subject.
The nominative Theme is the subject of an
embedded small clause.
There is a similarity between the synthetic and There is a variation between the synthetic and
periphrastic forms.
periphrastic forms.
Inversion with these verbs is ungrammatical except They are morphologically complex since they
for topicalization purposes.
consist of an ergative matrix V, ~ posse~~ive empty
verb and a small clause complement of a verb.
Table 3.
Mulder (1992) s- psych verbs and a-psych verbs
Mulder argues that all s-psych verbs are VPs which are multi headed. He
observes in Dutch that s-psych verbs have a reflexive counterpart, they can be agentive as
in (1), reflexive as in (2) and causative as in (3):
49
(1) Teun interested the student for English. (agentive in Dutch)
(2) The student interested for English. (reflexive in Dutch)
(3) Linguistics interested the student the most. (causative in Dutch)
Mulder claims that in constructions like (1) there is the corporation of either a
psych N or an ADJ into a causative V. In (2) the Experiencer argument is a "weak
reflexive" and in (3) there is a reflexive interpretation of a PP (see that in this case the PP
can not be realized overtly).
According to Mulder, the Experiencer argument is an argument who .either
possesses a mental state or is the subject of a psychological state (1992:122). However, this
possession that we see between an Experiencer and a psych noun in psych verbs is
inalienable possession. The Experiencer possesses the mental state which is represented by
the psych N. It is in fact an obligatory complement but since it is an inalienable possession
it is non-overt.
I. 2. 1. 5. Hale and Keyser (2002) on Psych Verbs
Hale and Keyser argue that, as it is the case with many transitive verbs,
Experiencer Subject psych verbs can not participate in middle constructions. The following
are their examples for that (14):
(1) *John's talent envies easily. (Everyone envies John's talent.)
(2) *French films love easily. (My kids love French films easily.)
On the other hand Experiencer Object psych verbs can form middle
constructions as in:
(3) Politicians anger easily. (The truth anger politicians.)
50
(4) This colt frightens easily. (Loud noises frighten this colt.)
(5) I worry easily. (Economic down-turns worry me.)
(6) Children bore easily. (Adult talk bores children.)
One of the reasons for Experiencer subject psych verbs' not forming middles
may be that the Experiencer is not affected by the action denoted by these verbs. In
Experiencer Object verbs it is affected. Therefore the first group does not meet the
Affectedness Requirement. Another explanation for these structures would be hidden in the
semantic connection between the internal and external arguments. In the following
examples, the psych nominals contain genitive nominals which refer to the Experiencer:
(7) Mary has my respect. (I respect Mary.)
(8) She has the boss's esteem. (The boss esteems her.)
(9) He has his children's love. (His children love him.)
Hale and Keyser claim that without these genitives, the Experiencer disappears
as in Mary has the respect or He has love.
The paraphrases of the sentences above would be:
(10)
I give my respect to Mary.
(11)
The boss gives her his esteem.
(12)
His children give him their love.
Hale and Keyser assume that psych nouns are bare nouns which have a part
relation to some whole entity. The following paraphrased examples show that the emotion
(here love) is possessed by the external argument (Mary) not by the internal argument (the
children) as it is the case with all Experiencer psych v~rbs:
( 13)
Mary loves her children.
(14)
~gives
her children her love.
51
This anaphoric properties of psych nouns are regarded by Hale and Keyser as
the reason of the failure of Experiencer subject psych verbs to form middles because
middle forms lack an external argument.
I. 2. 1. 6. Ginnis (nd.) on the Morphological Restrictions in Experiencer
Predicates
Basing herself on Pesetsky (1990) Causer Target/ Subject Matter distinction,
Ginnis gives examples from psych predicates which allow either an Experiencer or a
Suggestor subject. She describes the Experiencer as the argument which undergoes the
psychological state and the Suggestor as the argument which refers to the behavior
described by the psychological state and claims that a predicate can not have both a
Suggestor and T/SM arguments at the same time:
(1) Carol was fearful of earthquakes.
Expei'iencer
T/SM
(2) Carol's expression was fearful.
Suggestor
(3)
* Carol's expression was fearful of earthquakes.
Suggestor
T/SM
Ginnis explains the ill formedness of (1) with synthetic causatives like frighten
with a morphological restriction on causative elements.
In English, if possible, lexical causatives are inserted but if the verb has no
synthetic causative, the default form is used. Ginnis states that the T/SM violation with
synthetic causatives in English can be eliminated by using an "analytic default". These
processes that she mentions can be summarized for English like this:
52
(Causative)
<->
-ify
in env. Root+ ASP+
X: terr-, horr-
(Causative)
<->
-en
in env. Root+ASP+_ _ Y: fright ...
(Causative)
<->
(Causative)
<->
in env. Root+ASP+_ _ Z: please, disgust, anger ...
make
elsewhere
The aspectual head (ASP) in her analysis is not recursive but rather the VP is
recursive; one causative VP may take another causative VP as its complement, so, many
languages like Turkish have multiple causative constructions.
I. 2. 1. 7. Kural (1996) on Elementary Predicates and Psych Verbs in
Turkish
Kural ( 1996) identifies seven abstract elementary predicates for the syntactic
representation which are:
1.
2.
3.
4.
5.
6.
7.
Actor
Causative
Passive
Volitionality
Control
Inchoative
Beneficiary
ACT
CAUSE
PASS
VOL
CONT
INCH
BEN
Kural argues that there is a correlation between the Beneficiary and the
Experiencer roles. He sees the Experiencer role as an "instantiation" of the Beneficiary
role. He accepts the for PP in the specifier position of BEN as the Experiencer argument
which experiences the mentioned benefit. He states that John in the following sentences is
the Experiencer argument:
(1) The recent changes in the regulations benefited John.
(2) John benefited from the recent changes in the regulations.
53
The alternation above can also be seen m pairs like fear/frighten and like/
please:
(3) The recent changes in the regulations worried John.
(4) John worried about the recent changes in the regulations.
Kurallists four basic thematic roles:
1. Actor: manipulates the event
2. Experiencer: internally responds to the event.
3. Neutral: remains outside the event.
4. Patient: undergoes a change of state as a result of the event.
He introduces two other phenomena inherent in these roles; affectedness and
protagonism:
Affectedness
related to the change of state
actors and Neutrals are unaffected
Experiencers and Patients are affected
created arguments are not affected
destructed arguments are affected
only the affected objects can possessivize
affected arguments are allowed as the derived
subjects of middles but unaffected arguments do
not
inflectional materials like modality and negation
are irrelevant for affectedness
contextual information IS irrelevant for
affectedness
Table 4.
.Protagonism
related to active participation
protagonists actively determines the course of the
events
actors and Experiencers are protagonists
Neutrals and Patients are nonprotagonists
Kural (1996) Affectedness and Protogomsm
These two properties can be combined with four basic thematic roles in the
following way:
Actor
Experiencer
Neutral
Patient
Unaffected
Affected
Unaffected
Affected
Table 5.
Protagonist
Protagonist
Non-protagonist
Non-protagonist
Kural (1996) Affectedness, Protogomsm and Thematic roles
54
Kural mentions two hierarchies about the mapping of these thematic arguments
on to syntactic positions (1996:39):
1. Affectedness Hierarchy: An unaffected role is mapped onto a higher
argument.
2. Protagonism Hierarchy: A protagonist role is mapped onto a higher
argument.
He identifies ten possible thematic combinations for transitive dyadic verbs
(Kural, 1996:40). Among these, the ones that include the Experiencer argument can be
summarized in the following chart:
Few in number
Unaffected subject and affected object
-··· Both are protagonists
With verbs like con, influence, irritate, ridicule, stimulate, train etc.
NeutralCauses a clash between the A and P hierarchies.
Experlencer
The choice of subject depends on alternating individual verbs
admire, enry, fear, hate, like, pity, resent, believe, hope, think, want, understand, need,
taste take Experiencer subjects
amuse, confuse, disgust, frighten, offend, please, satisfy take Neutral subjects
Mostly
take Experiencer subjects
~ ~-E±0
Experience~;;;;:. ; ;~ A gap in the paradigm
Experiencer ;. ·
Two participants can not internally respond to the event
Lack an independent event which would be responded.
· Experiencer-; ~ ··· ~ A gap in the paradigm
Patient
An Experiencer subject can not change the Patient's state
ActorExpe_r.ien~~r -· ..
4
Table 6.
Kural (1996) Thematic Combmatlons Wlth the Expenencer Argument
Kural states that having a desire or thought is also a kind of change of state
because state is not only something physical but also something mental or social. An
Experiencer is an affected argument because it changes its mental attributes in some or
other way. An Experiencer is a protagonist at the same time because " ... the person who is
enjoying or wanting something is engaged in a mental activity that determines the enjoying
or the wanting" and an Experiencer mentally processes the event " ... a person who is
reminiscing or fantasizing about some situation must also be mentally constructing that
situation" (Kural, 1996:38).
55
This assumption of Kural is criticized by Dikken ( 1996) claiming that such a
sentence like:
(5) Sue believes in Santa Claus.
" .. .is perfectly true even Sue has always believed Santa Claus and will continue to believe
in Santa Claus till the end of time" (Dikken, 1996: 12) implying no change of state of mind
of Sue at all.
Kural's reason for the absence of Experiencer- Experiencer verbs is that the
Experiencer Subject makes the verb stative but Experiencer Object forces the verb to be
eventive and the result is a clash between the two.
Neutral- Experiencer verbs are said to cause a clash between the two
hierarchies because affectedness hierarchy says that the Neutral role should be mapped
onto the higher argument, on the other hand, protagonism hierarchy says that the
Experiencer role should be mapped onto the higher argument.
His observation is that in most of the world's languages the Experiencer
argument occupies the subject position. Therefore we can say that there is a precedence
relationship between the two hierarchies. Kural says that in this relationship the
protagonism hierarchy takes the precedence. It links the Experiencer to the subject
position. Only when protagonism can not determine, the affectedness hierarchy applies.
Kural shows this precedence rule as:
Protagonism Hierarchy> Affectedness Hierarchy
He claims that Experiencer Subject Neutral Object verbs like fear can be
analyzed with a simple VP structure like:
56
VP
A
DP
V'
/"'....
V
Experiencer like
DP
NEUTRAL
Figure 7. Kural (1996) Experiencer Subject Neutral Object Verbs
Neutral Subject Experiencer Object verbs have a different representation. Kural
argues that they are triadic verbs and there is a CAUSE which adds a Neutral argument:
VP
pp
V'
A
NEUTRAL
v
VP
~
CAUSE DP
V'
A
v
pp
I
I
please NEUTRAL
Figure 8. Kural (1996) Neutral Subject Experiencer Object Verbs
Here the Neutral argument in the specifier position of CAUSE initiates the
mental state of the Experiencer just like Pesetsky's Subject Matter and the Neutral
argument in the specifier position of the verb is just like Pesetsky's Target of Emotion.
Kural quotes Pesetsky's claim that the Causer argument is the Target when there is not a
Neutral argument as in:
(6) Suna Mehmet'i bahktan igrendirdi.
57
If there was not an internal argument like balzktan the Causer in this sentence
would be the Target Suna.
Kural takes all these as the evidence that this constraint is just related to surface
structure because these two arguments can co occur when they are kept apart in the surface
structure as in the periphrastic causatives in (7) and in verb particle constructions as in (8):
(7) Bill made Mary worry about his progress in the class.
(8) The lectures turned Bill on to classical music.
Kural explains the difference between the ungrammatical (1 0) and the
grammatical ( 11) as the type of emotion that they denote:
(9)
Mehmet Suna'nm kopeginden korktu.
(10)
*Suna Mehmet'i kopeginden korkuttu.
(11)
Ahmet Ay~e'ye caz1 sevdirdi.
His generalization is that psych verbs which refer to temporary emotions (such
as kederlen-, kzz-, sevin-,
~a~zr-,
utan-, iirk- etc.) obey this restriction while verbs which
refer to permanent emotions (such as bez-, bunal-, gocun-, igren-, imren-,
ku~kulan-,
oykiin-, sev- etc.) do not. In Koral's analysis the verb fear has a different status among
these verbs because fear can be both a permanent property towards an object or it may be a
temporary state "aroused anew in each situation" (1996:133).
I. 2. 2.
Semantic and Aspectual Approaches to Psych Verbs
In this section, first the aspectual approaches to psych verbs will be examined
through the analyses of Grimshaw (1990), Tenny (1994), Voorst (1992), Filip (1996), Arad
(nd.) and Levin and Hovav (2002). Then, other significant semantic approaches based on
58
non aspectual criteria will be overviewed through the analyses of Dowty (1991) and
Hatory (1997). Wechsler (1995), Krifka (2001) and Kordoni (2000 and 2001) focus on the
semantic concepts of notion and underlying, and, Dabrowska (1994) examines the case
marking of Experiencer arguments in Polish.
I. 2. 2. 1. Grimshaw (1990) Prominence Theory and Psych Verbs
Grimshaw (1990) assumes that the argument structure has its own internal
structure or organization and this organization is a reflection of the lexical semantics of the
verb. Thus, the argument structure of a verb can be predictable from the general
characteristics of the meaning of a verb. This assumption commits itself to the idea that the
syntactic representation can be derived from the semantic representation.
She claims that there are prominence relations among the arguments of a given
verb and, the argument structure is a representation of these prominence relations. These
prominence relations are determined by two properties of verbs which are the thematic
properties and the aspectual properties. The theory presents the
ex~ernal
argument as the
most prominent argument in thematic and aspectual dimensions which is equal to the dstructure subject. An argument is accepted to be either external or internal according to its
relation to other arguments in the argument structure. The theory assumes the following
hierarchy for the arguments:
(Agent (Experiencer (Goal/Source/Location (Theme))))
The hierarchy indicates that the Agent is the highest argument, and then comes
the Experiencer, Goal, Source, Location and lastly the Theme. The most prominent
59
argument in the argument structure always occupies syntactically the most prominent
position.
In this theory, the special status of psychological verbs is due to the mismatch
between their thematic and aspectual properties. Fear class verbs and their prominence
relations are like other agentive verbs. Frighten class verbs on the other hand do not meet
expectations since their thematically the most prominent argument (Experiencer) does not
occupy the syntactic subject position:
Fear
(x
Exp.
(y))
Theme
Frighten (x
(y))
Theme Exp.
Unlike Belletti and Rizzi (1988), Grimshaw ( 1990) argues that the Theme of a
frighten verb is a d-structure subject. The idea defended in the theory is that these two
classes of verbs have the same thematic prominence relations although they differ in the dstructure representation of their arguments. The mismatch is because of the difference in
their aspectual properties. More specifically, frighten class verbs are causative but not
stative while fear class verbs are stative, and do not have an event reading.
The result is that the causal status of an argument plays a role for this
argument's syntactic realization either as a subject or as an object. Grimshaw regards this
dimension as autonomous from the thematic dimension. She mentions two different kinds
of dimensions; one thematic, the other causal. "Each of the two hierarchies imposes its
own set of prominence relations on this collection of arguments ... " (1990:24). The causal
hierarchy is:
60
(Cause (Other (. .. )))
For frighten verbs, the first element in the thematic hierarchy corresponds to the
second position in the causal hierarchy and vice versa.
For the fear class, the Experiencer argument, both thematically and aspectually
the most prominent one but for frighten class the Theme is the aspectually the most
prominent.
Grimshaw states that frighten type verbs cross linguistically seem to be more
stable than the fear type. She defines another class of psych verbs whose semantics are
similar to fear type verbs which can be regarded as unaccusatives (such as concern and
please).
The underlying idea of the theory is that every argument has a prominence
status relative to the other arguments in the mentioned two dimensions. The external
argument is the one which is the most prominent in two dimensions. A verb can only have
one external argument since only one argument can have maximal prominence. Frighten
verbs are different from other causatives in that they have no external arguments.
Therefore, they can not undergo such processes like nominalization or passivization which
require the suppression of the external argument. Here, Grimshaw distinguishes between
the two notions; the d-structure subject and the external argument. Frighten verbs have d- ·
structure objects but not external arguments.
She also differentiates between causatives and unaccusatives. She states that
frighten verbs have an argument which has a maximal
~pectual
prominence but not an
external argument which has a maximal thematic prominence. The unaccusatives too, do
not have an external argument since they do not have a first sub event. She claims that the
only argument of unaccusative verbs is in the d-structure object not a subject. Although
61
both classes of verbs i.e. unaccusatives and frighten type verbs have internal arguments,
Grimshaw says that the aspectual status of these arguments are different from each other.
I. 2. 2. 2. Tenny (1994) Aspectual Interface Hypothesis and Psych Verbs
According to Tenny (1994), the intersection or the visible side of the lexical
semantics is the side of the aspectual properties which regulate much of the linking. These
aspectual properties which are related to syntax are of a very limited kind. Specifically she
sees delimitedness as the only aspectual property which plays a role in the syntax
semantics interface. For her, delimitedness refers to an inherent endpoint of events. It
depends not only on the verb but also the other elements in the verb phrase.
She argues that the internal arguments of stative verbs do not undergo any
change or motion. They do not also have an internal temporal structure. They just describe
states.
., She demonstrates the difference between the fear and the frighten type verbs
with the differences between their entailments as in (1) and (2):
(1)a. The truth frightened John in five minutes.
Entails that
b. It took five minutes for John to become frightened.
c. John was frightened at the end of five minutes.
(2) a. ?John feared the truth in five minutes.
Does not entail
b. It took five minutes for John to fear the truth.
c. John was feared at the end of five minutes.
62
She gives further evidence for the differences between these two classes by
using rate adverbials to modify the argument that undergoes a change of state. In this way
we can measure out the change of the changed argument as in:
(3) The truth interested John only slowly, since he was not imaginative by
nature and was slow to see its implications.
A delimiting expression can also be used with a causative change of state verb
to refer to the ends tate of the internal argument as in:
(4) The movie frightened the children to death.
(5) *The movie feared the children to death.
Tenny describes the constraint related to the example (9) as the Measure-Out
Constraint which states that only the direct internal arguments (not external arguments)
can undergo a change of state and only they are able to measure out an event. The verbs
whose Experiencers are the internal arguments can express a change of state but verbs
whose Themes are the internal arguments can not. Only the verbs which have Theme
subjects can be causative while Experiencer subject verbs can not (Tenny, 1994:67).
Tenny's argument related to psych verbs is that they conform to the aspectual
restrictions on the internal and external arguments in the Aspectual Interface Hypothesis.
Their internal arguments are also constrained by the Measuring-Out Constraint.
She
simply sees the Experiencer argument of a psych verb as an event participant which
sometimes measure out and sometimes does not measure out events. As an internal
argument, it measures out the event, as an external argument, it does not. The Theme
argument of a psych verb, on the other hand, undergoes a change of state. If the Theme is
the internal argument, the verb (Experiencer- Subject Theme- Object verb) is not a change
of state verb; otherwise it is (67).
63
I. 2. 2. 3. Voorst (1992) and the Aspectual Semantics of Psych Verbs
Most of the theories about the aspectual properties of psych verbs including
Grimshaw and Tenny claim that the differences between the classes of psych verbs and
peculiar properties of the verbs offrighten class are due to the differences in their aspectual
properties. Grimshaw sees fear type verbs as stative and frighten type verbs as causative.
Unlike this common assumption, Voorst (1992) argues that there are not aspectual
differences in the deep structures of these classes and therefore, the thing to do is to find
out descriptions other than aspectual and thematic ones.
Voorst identifies four classes of psychological verbs in English and claim that
even though there are some semantic differences among these classes; all of them behave
similarly in the aspectual tests. His classification of psych verbs ( 1992:66, 67) can be
summarized like this:
1. Class 1 Psych Verbs: Psychological uses of action verbs:
He struck me as rather odd.
2. Class 2 Psych Verbs: Psychological verbs with an intentional subject:
The clown tried to amuse me.
3. Class 3 Psych Verbs: Psychological verbs with a non-intentional subject:
These experiences amused me tremendously.
, 4. Class 4 Psych Verbs: Psychological verbs of dislike type in which the
subject is the centre of the psychological experience.
These art-connoisseurs admire Van Gogh.
Voorst in the analysis of activity and accomplishment verbs reaches the
following conclusions about psych verbs:
64
•
They do not allow end reading. Begin reading is also sometimes
unacceptable since it is difficult to predict the exact moment of the
occurrence of any kind of feelings.
•
They allow the adverb for X minutes similar to activities and states.
•
With the use of the adverb almost, all classes of psych verbs have the
interpretation that the event has failed to occur.
•
Unlike activities and accomplishments, individuation of the direct object of
psych verbs does not affect their aspectual semantics.
•
They do not imply a process leading to a change of state.
Voorst then, analyzes the behavior of psych verbs comparing them to the
achievement and state verbs and observes that they (just like achievements and states) refer
to the beginning of the state or event. With the adverb in X minutes, only a begin reading is
possible for them.
Voorst distinguishes between the states and events analyzing the behavior of
psych verbs and observes that:
•
They describe a more permanent feeling without the progressive form. On
the contrary, when the verb is in progressive, a less permanent feeling is ..
conveyed.
•
When their subjects can not be interpreted as agentive, these verbs are
states.
•
Their imperative form is better with an optative meaning and with an
implied agentive subject rather than giving an order.
•
They allow passivization unlike states.
•
Unlike states, they allow adverbs which specify the intensity.
65
As a result, psychological verbs are not like accomplishment verbs because they
do not imply a process. They are not also like activities because the individuation does not
transform them into accomplishments. Lastly, they are not like states because they take
place or happen unlike them. These all leads him to conclude that psychological verbs are
achievements.
Another observation of Voorst is that manner adverbs such as carefully and
slowly can be used with psych verbs with perfect grammaticality but these adverbs do not
modify the psychological process. For example when we say:
(1) Peter carefully frightened the grizzly bear to chase it out of its backyard.
carefully does not refer to the performing the process of frightening.
Another important distinction is between subjects denoting individuals and
subjects denoting events. In (2) the subject is the individual himself while in (3) the subject
is not the individuals but something about the individuals:
(2) He demonstrated that the situation was getting worse each day.
(3) These kids demonstrated that the situation was getting worse each day. (the
kids' starving appearance etc.)
The measurability that Voorst mentions is not the kind of measurability that of
Tenny's. Voorst puts forward that psychological verbs are unmeasurable events since we
can not talk about the existence of such a process which leads to final state. That means,
just like achievement verbs, psych verbs do not have a culminationpoint.
Voorst adapts Ryle's (1949) classification of emotions expressed by psych
constructions. He mentions three types of emotions:
1.
Inclinations: They refer to permanent emotional dispositions or
traits of character.
66
Moods: They are less permanent dispositions or "short term
2.
tendency words".
Agitations: They are more or less lasting dispositions.
3.
Voorst claims that "Constructions with psychological verbs are centered around
moods and those containing psychological uses of accomplishments around inclinations
underlying moods" (1992:90). For him, the reason for psychological verbs' not being
accomplishments is that they are centered around moods which can be manipulated.
I. 2. 2. 4. Filip (1996) on the Psych Verbs in Czech
Filip challenges the principles in the Aspectual Interface Hypothesis giving
evidence from psych verbs in Czech and English that telicity or measuring out the events
can not be the basis for the syntax semantics interface.
Filip argues that only the psychological predicates which are achievements (the
ones which refer to spontaneous changes of mental state) are telic. He gives examples from
some psychological causatives (like calm, disillusion, sadden, soothe, disarm etc.) which
do not entail a clear cut final stage. These verbs do not co occur with adverbs like halfway
and therefore they are atelic as in:
(1)
?The music halfway saddened John.
(2)
*The high-pitched noise halfway distracted her.
Filip argues against Tenny's test with incremental expressions like a little bit,
saying that this expression relates only events as a whole not their increments or parts. For
example when we say:
(3)
The music saddened John a little bit.
67
we do not mean the part of a larger event but we mean a kind of lower intensity of the
whole psychological event.
In Czech Filip identifies three types of psych verbs:
1. Nominative Experiencer
2. Accusative Experiencer
3. Dative Experiencer
According to Filip there 1s a correlation between the case assignment and
thematic roles:
•
Nominative case of the Experiencer is associated with the existence of
sentience and volition
•
Accusative or dative case of the Experiencer is associated with the
existence of a lack or lower degree of control
•
Nominative Experiencers are either the cause of the event or they are
volitional.
•
Accusative Experiencers are either causally affected or undergo a
change
of state.
Accusative
Experiencer
verbs
can
undergo
passivization.
•
Dative Experiencer verbs can not undergo passivization.
•
Dative Experiencers are generally beneficiaries.
•
Dative Experiencer verbs can not be modified with manner adverbs like
in a pleasant manner, passionately, bitterly but they can be modified
with degree or intensity adverbials like a lot and a little.
To conclude, Filip's main argument is that telicity or any other semantic
property is not useful in explaining the linking properties of psych verbs in Czech.
68
I. 2. 2. 5.
Arad (nd.) on Psych Verbs
Arad (nd.) has two basic claims about the syntactic and semantic properties of
psych verbs. The first one is that the different features that belong to Object-Experiencer
(OE) verbs are only seen in their stative reading i.e. with an agentive reading these features
do no exist. Her second claim is that OE psych verbs are not inherently psych but they are
either formed from normal predicates or they do have also non-psych senses. The main
argument is that the peculiar features of OE psych verbs belong to their stativity.
Arad shows that OE psych verbs have three readings which associate with
different syntactic constructions. This shows that a great number of verbs have both a
psych and a non-psych sense and even these senses have different readings in themselves.
The mentioned three readings are:
1. Agentive reading: There is an intentional Agent and a change of state in the
Experiencer:
(1) Nina frightened Laura (to make her go away).
2. Eventive reading~ There is an unintentional Agent and a change of state:
(2) The explosion/ noise/ thunderstorm frightened Laura.
3. Stative reading:
- There is neither an Agent nor a change of state in the Experiencer.
- The Experiencer experiences this mental state as long as she perceives the
Stimulus.
- There is a kind of stative causation (unlike the other two readings in which
there is an active causation).
- The stative Causer is not an external argument
69
-The stative Causer is not an affected argument.
-The stative Causer is an "external internal, argument (Arad, nd: 16).
-The stative Causer can only be generated under the specifier position ofVP.
Arad notes that there are some verbs which allow only one of these readings.
For example while the verb frighten allows three readings, verbs like concern and worry
allow only stative readings. Arad's main prediction is that all of the psych properties exist
only in the stative reading of OE verbs. In the agentive reading these verbs are like normal
transitive verbs.
Another claim is that almost every argument position can be interpreted as an
Experiencer. It can be a direct object, an indirect object or a PP. What determines the
interpretation of an argument as an Experiencer or a Goal etc. is the verb and its arguments
(Arad, nd: 13- 20).
Arad adopts Bouchard's (1995) assumption that all verbs can be interpreted as a
psych verb if some conditions are met. For this, first there should be an animate argument
which can be interpreted as an Experiencer and second, there should be an internal
argument which can be interpreted as an emotion or a mental state (Arad, nd: 14).
I. 2. 2. 6. Levin and Hovav (2002) on Psych Verbs
Levin and Hovav (2002) claim that to decide whether an event with two sub
events is a complex event or not, we should look at the temporal relations between the subevents. Therefore the internal temporal constitution of an event which is an aspectual
feature is relevant to argument expression.
70
They claim that mental verbs like read, copy, memorize, translate, study, recite
and perform etc. are like accomplishments but they are not complex events. The object that
is read is not affected by the reading activity. The reader in a way internalizes the text and
gains a mental representation from it. They argue that these verbs have incremental objects
like verbs of consumption rather than verbs of change of state. Their objects are
measurable NPs, thus they are telic.
They analyze the mental verb read and say that it has two sub events. The first
sub event is scanning the text and the second sub event is forming a mental representation.
The point is that these two sub events may be temporally dependant because while the text
is scanned, a mental representation is formed at the same time. Thus they do not regard
these verbs as complex events (Levin and Hovav, 2002:15).
They further point out that incremental Themes of these verbs do not have to be
syntactically expressed as the direct object as in (1):
(1) Kelly read/studied.
(2) Kelly read/studied from the textbook.
I. 2. 2. 7. Dowty (1991) on Proto Roles and Psych Verbs
Dowty ( 1991) mentions about an indeterminacy in argument selection with
some verbs which he calls lexical doublets. These verbs have different argument
configurations but they express the same relation. For example, verbs like buy and sell,
borrow and lend do not distinguish between their buyer and seller, borrower and lender
with respect to the properties of proto roles oftheir arguments.
71
Under his view, psychological predicates are also different kinds of doublets.
He classifies the following verbs(Dowty, 1991:579) into two groups as:
Experiencer-Subject
Stimulus-Subject
x likes y
x fears y
x supposes (that) S
x regards y (as) VP
x is surprised at y
x is disturbed at y
y pleases x
y frightens x
(it) seems (to) x (that) S
y strikes x (as) VP
y surprises x
y disturbs x
In the first group of verbs, Experiencer percepts the Stimulus therefore the
Experiencer is "sentient/ perceiving" but in the second group Stimulus argument is the
Causer of the "emotional reaction or cognitive judgment" of the Experiencer. Dowty
argues that both of these arguments have " ... a weak but apparently equal claim to
subjecthood" (1991 :579).
The first group of verbs are stative verbs but verbs in the second group are
sometimes stative sometimes inchoative. The Experiencer has the proto patient properties
when there is an inchoative interpretation but this is not the case for the statives. Therefore,
according to Dowty's argument selection principles, this argument has the most proto
patient properties (although both arguments have the same amount of proto agent
properties) and it becomes the direct object.
Dowty in his proto role analysis of psychological verbs claims that the
Experiencer Object verbs have a Stimulus/Causer argument which causes a cognitive
judgment or emotional reaction in the Experiencer. The Experiencer argument is seen as
the participant who undergoes a kind of change of state. He notes that the Experiencer
argument has the perception of the Stimulus. This property is a proto Agent property for
the Experiencer. The Stimulus on the other hand causes the Experiencer to perceive it
72
(which is also a proto Agent property for the Stimulus). This is which causes a conflicting
mappmg.
I. 2. 2. 8. Hatory (1997) and the LCSs of Psych Verbs
Hatori (1997) views psych verbs as ordinary transitive verbs in terms of -able,
-er and middle formation as in the following examples:
(1) Sue frightened easily. (middle formation)
(2) excitable, irritable, upsettable (-able adjective formation)
(3) disturber, enchanter, flatterer, startler (-er noun formation)
Hatori observes that some verbs of change of location can be extended to more
abstract semantic fields such as a change of state; therefore they both have the same LCSs
such as the following ones:
(4) a. John drove his mother to the station.
b. John drove his mother mad.
(5) a. The breeze moved the leaves slightly.
b. The story of their sufferings moved us deeply.
(6) a. The branches touch the roof.
b. Her kindness touched me profoundly.
Hatori observes in these examples that when these verbs are used with their
psychological meaning they co-occur with such adverbs like deeply, profoundly etc. and
they are also able to derive psych adjectives with -ing as in moving and touching.
73
Hatori suggests two approaches to analyze psych verbs. The first approach sees
the Experiencer argument of these verbs as the Location or the Goal. The second approach
sees it as the Theme.
According to the first approach the Experiencer can be either in the dative or in
the oblique case. The examples in which the preposition to coming after some psych
adjectives are presented as the evidence for that:
(7) The point was clear to me.
(8) They are interesting to the students of comparative literature.
The second approach that regards the Experiencer argument as the Theme sees
us as the Theme and film and the director as the Agent:
(9)
The film frightened us.
( 10)
The director sent us.
Another observation of Hatori is that psych verbs can co-occur with resultative
expresswns:
(11)
He awed them into obedience.
(12)
He was shamed out ofbad habits.
(13)
His suffering distressed him into committing suicide.
Hatori's argument is that the verbs that are compatible with path expressions
have the LCS with either GO or MOVE function. However, some other psych verbs can
not co-occur with path expressions:
(14)
* The storyteller's jokes amused the children into giggling.
Hatori claims that in the LCS of these verbs, there is not a GO or MOVE
function but a BE function, therefore, they do not refer to an emotional motion or an
Experiencer that undergoes it. They have either a stative or an inchoative meaning. With
74
these verbs the Theme argument can be realized as a with phrase mostly in passive
constructions:
(15)
The Chinese dinner satisfied Bill.
(16)
Bill was satisfied with the Chinese dinner.
He also gives examples from a few number of psych verbs which can be
followed by both a with phrase and a path phrase. One of them is the verb bore and the
verb worry:
(17)
... a handsome engineer who bored me to tears with his tales of
motorway maintenance.
( 18)
You worry me to death with your chatting.
To sum up, Hatory's analysis of LCSs of EO psych verbs classifies them into
two as one with a GO/ Move function and the other with a BE function.
There are some other psych verbs which are used with body parts. Here, the use
of body part expressions (like head, mind, nerve, head, soul, stomach and chest etc.) gives
a kind of metaphorical meaning. The observation is that identical body part nouns are used
across languages. In a number of languages these body part expressions can be used as the
Experiencer argument. This should be because humans conceptualize emotions very
similarly cross linguistically.
I. 2. 2. 9. Wechsler (1995) Notion Rule and Psych Verbs
Wechsler (1995) introduces a concept called notion to define the semantic
constraints on argument structure. He claims that this concept underlies the semantic
75
constraints even of psychological verbs which constitute a problematic class. He defines
notions as mental entities or concrete cognitive structures (Wechsler, 1995:32).
In order to explain the semantic constraints on argument structure, he defines a
set of notion rules. He analyzes the verbs in the following sentences (1), (2), (3) and (4):
(1) a. John wants the cat.
b. John has a notion of the cat.
c.
* The cat has a notion of John.
(2) a. John likes Mary.
b. John has a notion of Mary.
c.
* Mary has a notion of John.
(3) a. John fears Mary.
b. John has a notion of Mary.
c. * Mary has a notion of John.
(4) a. John is expecting Fred.
b. John has a notion of Fred.
c.
* Fred has a notion of John.
He formulates the following Notion Rule for the psych verbs want, like, fear
and expect in (1), (2), (3) and (4);
The individual (here John) has a notion of X (the object I content of his mental
state) but this X does not have to have a notion of John necessarily. In each of the
sentences the (b)s are entailed but the (c)s are not. Thus reverse entailments are not true.
For example in (1), in order to want the cat, John should have a notion of the
cat but that does not mean that the cat should also have the notion of John.
Wechsler claims that this notion rule is also valid for the verbs of perception:
76
(5) John saw the cat.
a. John has a notion of cat.
b. *The cat has a notion of John.
Wechsler claims that frighten type verbs seem to violate the notion rule but they
do not. They, like the fear type, are consistent with the notion rule. He emphasizes that
frighten type verbs are eventive. They describe the causation of the mental state but they
do not describe anything about the intentionality of the result state.
He distinguishes between Cause and Content saying that they are distinct
concepts. However they may sometimes refer to the same entity as in:
(6) The dog frightened the baby.
Wechsler accepts the existence of some psych verbs as counter examples to the
notion rule such as concern and preoccupy:
(7) *Toxic waste concerns the Senator deeply- he just happens to be unaware of
its existence.
Wechsler says that for an X to concern or preoccupy a Y, Y must have some
notion of x. He argues that these verbs are few in number, their number does not seem to
increase and they are problematic for language learning.
I. 2. 2. 10.
Kritka (2001) on Psych Verbs
According to Croft (1991) one of the tests to distinguish between the
Experiencer- Subject and Stimulus- Subject verbs is that only the second type of verbs
allows by phrases. However Experiencer- Subject verbs use causal clauses instead of byphrases as in (Krifka, 2001:11):
77
(1) John pleased his boss by coming in early everyday.
(2) a. *John was liked by his boss by coming in early everyday.
b. John was liked by his boss because he came in early everyday.
Other than the known two groups, Krifka identifies two more classes of
psychological verbs. The first group does not have a direct object but a Stimulus subject:
(3) The painting appealed to Mary.
The second group does not have a direct object too but has an Experiencer
subject:
(4) The soldiers rejoiced about the victory.
Krifka claims that Experiencer- Stimulus verbs should also include perception
verbs such as see verbs like feel, hear, smell, taste, notice etc. , sight verbs like discover,
examine, inspect, perceive and verbs that refer to propositional attitudes like know, believe,
doubt etc.
Krifka exemplifies that although Stimulus- Subject verbs refer to punctual
changes, they can be used with the progressive tense:
(5) The storm was frightening the people.
but Experiencer-Subject verbs have a stative reading and can not be used with progressive:
(6) *The people were fearing the storm.
Krifka claims that although most of the Stimulus- Subject verbs in languages
are often regarded as causatives, the kind of causation by a Stimulus is not actually
causation since it does not undergo causative alternation as non psych verbs do:
(7) a. Someone broke the glass.
b. The glass broke.
(8) a. Someone frightened John.
78
b. *John frightened.
According to Dowty's analysis, if the verb refers to a change of state, the
Stimulus has a proto Agent property and it is realized as the syntactic subject. Similarly if
the Experiencer has a notion of the Stimulus, the Experiencer has a proto Agent property
and it is syntactically realized as the subject. What he means by notion is Wechsler's
(1995) concept of notion which simply refers to knowing the entity which is the Cause of
the experience.
Other differences between the Stimulus- Subject and Experiencer-Subject verbs
are listed by Krifka (200 1: 13-15) as follows:
•
Reciprocals reinforce the agentive reading of Stimulus-Subject verbs:
The children entertained us. (ambiguous between agentive and non
(9)
agentive readings)
(10)
•
The children entertained each other. (only the agentive reading)
In some languages such as Italian, Stimulus- Subject verbs can not be
reflexive (since their subjects are derived subjects as argued by Belletti and
Rizzi).
•
The object of Stimulus- Subject verbs can bind anaphors within the subject.
•
The subject has fewer subject properties than the object in Stimulus-Subject
verbs (this pattern assigns some subject properties to the Stimulus such as
syntactic position, verb agreement, case marking etc.).
•
Experiencer-Subject verbs are regarded as the basic form but the number of
Stimulus-Subject verbs is greater in (which is not the case in Turkish).
79
I. 2. 2. 11.
Kordoni (2000 and 2001) on the Psych Verbs in Modern
Greek
Kordoni, (2000) provides evidence against the idea that the Experiencer Object
psych verbs are causative constructions. Within the framework of Head Driven Phrase
Structure Grammar (HDPSG), she identifies three classes of psych verbs in Modem Greek
(MG).
The first class is similar to fear class in English. It includes verbs with an
Experiencer Subject and a Theme Object. The second class just like the frighten class in
English. It has a Theme or a Cause argument which is in the nominative case and which
agrees with the verb. There is one more class of psych verbs in Modem Greek. This class
of verbs has dative Experiencers and nominative Theme arguments.
Kordoni (2000) analyzes Modem Greek Experiencer- Subject ·psych verb
constructions (MG ESPVCs) in the framework of Wechsler's Notion Rule. One of the
claims that he proposed is the object arguments of MG ESPVCs are "semantically
underspecified" (199). By semantically underspecified he means, in these constructions,
the meaning of the verb does not pose any constraints over the object argument and it is
nonspecific about the nature of this argument. In the following examples (1) and (2) Gianis
has a notion of Mary in order to love or fear from him. However Mary can be either
cognitive or non-cognitive and we need to have some other contextual information to be
sure about that:
( 1) a. 0 Gianis agap a tin Maria.
John loves Mary.
b. John has a notion of Mary.
c. ?Mary has a notion of John.
80
(2) a. 0 Gianis fovate tin Maria.
John fears Mary.
b. John has a notion of Mary.
c. ?Mary has a notion of John.
The conclusion that Kordoni (2000) reaches is the subject NPs of ESPVCs
have a notion of the entity denoted by the object verb according to Wechsler's Notion
Rule. The non cognitive participants of MG ESPVCs are syntactically realized as the
object NP.
Dowty in his proto Agent and Patient property analyses also claimed that the
experienced (stimulus) argument of ESPVCs has neither proto Agent nor proto Patient
properties~
therefore he claimed that they are semantically underspecified. In this way
Dowty's and Kordoni's approaches are parallel to each other.
Grimshaw (1990) has also claimed that accusative Experiencer- Object psych
verb constructions have two different readings; agentive and psychological.
Agentive reading is possible when the nominative argument is animate. This
animate argument is interpreted as the deliberate and volitional Agent who causes the
object Experiencer to experience the mentioned emotional state. With this reading the
accusative Experiencer can not be doubled. In Greek, in such a sentence like (3):
(3) Giannis upsets Mary.
Giannis knows Mary (he has a notion of Mary), and Mary also has a notion of Giannis i.e.
the individual denoted by the subject NP has a notion of the individual denoted by the
object NP and vice versa.
When there is a causative reading the clitic on accusative Experiencer is
doubled obligatorily. The interpretation is there is something about Giannis that causes
81
Mary to experience an emotional state. This sort of constructions is similar to what
Grimshaw calls causative psych verb constructions as in (2):
(4) The thunderstorm frightened the child.
In both of these constructions Giannis and the thunderstorm do not have any
notion of the argument experiencing the mental state i.e. Mary and the child are neither
volitional nor intentional. They also do not participate in the mental state, they just cause it.
Thus, in these constructions only one of the participants is cognitive and the
other participant is just the Causer. Therefore, these constructions can be regarded as
normal causative constructions.
Grimshaw claims that ES psych predicates have their external arguments and
therefore they can undergo passivization. Kordoni (2001) gives counter examples for this
claim from MG ESPVCs. While (5) and (7) are grammatical in MG, (6) and (8) are not:
(5) His parents love John.
(6)* John is loved by his parents.
(7) Mary envied the neighbor's house.
(8) *The neighbor's house was envied by Mary.
He concludes that MG ESPVCs are neither passivized nor they are the passive
forms of the corresponding EOPVCs. The experienced (stimulus) argument of them are
realized either as the object ofthe sentence or the complement of a PP. The accusative NP
does not behave like a normal object since it can not be passivized.
82
Dabrowska (1996) on the Dative and Nominative
I. 2. 2. 12.
Experiencers in Polish
Dabrowska, (1994) in the analysis of the case marking of Experiencer
predicates in Polish argues that mental experiences have two aspects whose properties can
be summarized in the following table:
objective
conscious
public
active . . .
subjective
irrational
personal
passive ...
Table 7.
Dabrowska (1994) The two Aspects of Mental Experiences
Dabrowska claims that we have a tendency to identify ourselves with the first
aspect of mind rather than the second. Therefore, we tend to conceptualize ourselves as
active, conscious and rational etc. rather than passive, unconscious and irrational with the
effect of cultural values.
Dabrowska observes that some verbs in Polish have the Experiencer arguments
either in nominative or in dative and there are also some verbs which allow both cases. She
claims that semantic motivations behind the choice of case marking can be defined with
reference to the concept of "sphere of awareness". Some of the basic characteristics which
can be associated with these cases in her analysis can be summarized as follows:
Is defmite, concrete and active.
Is with a participant who has Agentive properties
Foregrounds the Experiencer's active role.
Backgrounds the sphere of awareness.
Is used with predicates which refer to a definite
desire, intention, determination or resolution.
Emphasizes volitionality and control.
Involves effort and action.
Table 8.
The participant is affected by a state or process
occurring in his personal sphere.
Foregrounds the sphere of awareness.
Depicts the Experiencer as passive.
Portrays the Experiencer as the passive receiver of
impulses.
Describes physiological drives and cravings which
come from the instinct.
Describes effortless, unplanned, can't help
situations.
Focuses on the irmer experiences of he Experiencer.
Dabrowska (1994) Nominative and Dative Experiencers
83
She accepts that these semantic properties do no not make all of the
grammatical properties predictable. She just claims that grammatical forms are motivated
by them. For her, the choice between the two cases" ... has the effect of highlighting some
aspects of the situation and hiding others" (1994:1047).
84
II.
ANALYSIS OF THE PSYCH VERBS IN TURKISH
II.
1. What is Psych?
This section presents a brief introduction to psych verbs in Turkish to clarify
the basic assumptions of the study with an initial analysis of some semantic and
morphological aspects of these verbs.
In the realm of psych verbs, there is a multitude of terms like mental, cognitive
and emotional. The first two terms are commonly used interchangeably to refer to verbs
connected with conscious mental processes such as to know, to learn and to understand.
The term emotional on the other hand is a quality of verbs connected with feelings such as
love, like, hate, fear etc. In some studies we see the verbs of one group under the title of
the other group or sometimes it is the case that these verbs are altogether regarded as
psych.
In fact it is not easy to distinguish between the two since mental and emotional
domains seem to be interconnected in our daily lives too. Some activities have emotional
bases as a point of initiation
as
opposed to other activities which have mental bases.
Accordingly, some of the verbs that we examined in Turkish belong to mental domain such
as algzla-, anzmsa-, anlamlandzr-, ayrzmla-, belle-, bilin9len-, bil-, degerlendir-, dii,Jiin-,
kap-, kavra-, kzvzr-, ogren-, yor- while some of them unquestionably correspond to the
emotional domain such as acz-, arzula-, bayzl-, begen-, bunal-, 9zldzr-, giiven-, igren-,
kaygdan-, kork-, onemse-, iimitlen-, iiz-, zevklen- and so on.
There are some other verbs which can be considered either as mental or
emotional. Although it is almost for sure that a verb like kavra- belongs to the mental
85
domain and a verb like begen- belongs to the emotional domain, a verb like ilgilen- can
either be mental (for example if you are interested in a subject or an idea) or emotional (for
example if you are interested in a boy or a girl to mean to be attracted to). It is again hard
to say whether benimse- "to adopt" occurs in the emotional or the mental domain.
Sa~zr
"to be surprised" seems to belong to the mental domain if taken as an act of being
surprised for something different from your existing thoughts related to this thing. If this
act of being surprised is for something that is different from what you emotionally
expected then the verb belongs to the emotional domain. The analysis of such verbs which
can be regarded as fuzzy cases shows that some of them are semantically closer to the
mental domain such as afalla-, aldan-, apl$-, ayart-, kan-, kur-,
ku~kulan-, ~ah
u9-,
yargzla-, zannet- and some of them are semantically closer to the emotional domain such
as avun-,
bagz~la-,
benimse-, ilgilen-, onemse- .
The mental and the emotional domains, first, as we stated are indistinguishable
and have some intersections, and second, verbs belonging to emotional and mental
domains have structural correspondences rather than differences. Therefore we have
enough reasons to regard both mental and emotional verbs as psych.
None of the studies in the psych verb literature offers a clear cut definition or at
least attempt to define what a psych verb is. In their appendices there are neither lists of
verbs that they accepted as psych nor a set of criteria specifying the features that these
verbs possess or do not possess but it seems that a kind of consensus exists as to what they
are. Therefore, before everything, we need to define what a psych verb is or at least what
we take it to be in this study.
In order to be a psych verb, a verb should:
86
1. be semantically related to either the emotional or the mental domains i.~. it
should describe an event occurring in one of these domains.
2. have
an
animate
Experiencer
argument
who
expenences
the
mentaVemotional state.
3. lead to a change in the psychological state of the Experiencer.
II. 1. 1.
The Experiencer Argument
It is obligatory for a psych verb to have an Experiencer argument.
Semantically, this argument experiences the men:taVemotional state that the verb describes.
This argument is regarded by Kural (1996) as an instance of the beneficiary argument.
The discussions on psych verbs are centered mostly on the structural positions
of the thematic arguments Experiencer and the Theme. The syntactic position of the
Experiencer argument is either the subject (mostly the subject) or the object position as it is
the case in Turkish:
(1) Ali Ay~e'yi seviyor.
Experiencer
(2) Ali Ayse'yi tiztiyor.
Experiencer
The surface case marking properties of the Experiencer argument may vary
according to the structural position that it occupies; it is nominative marked in the subject
position and accusative marked in the object position.
The Experiencer argument is almost always an animate being. It is hard to
think of any situations that an inanimate being experiences one of the mentaVemotional
states like aldzr-, r;eldn-, dan!-, duygulan-, hzrslan-, hisset-, hiiziinlen-, kaygzlan-, kopiir-,
kiis-, onemse-, pipiriklen-, sez-, szkzl-, iimitlen-, yadzrga-.
87
Therefore the emotional domain is one of the unique belongings of human
being and it is at least at that moment, not realistic to think of an inanimate being whose
psychological state changes or which understands implied meanings etc. Moreover, even in
such cases where the subject is not a nominal referring to an inanimate being, we tend to
interpret it as animate.
II. 1. 2.
The Theme Argument
Semantically the Theme role is the most neutral or inactive thematic role
among the other thematic roles, although it does not seem to be equally inactive in the
following examples (1) and (2) with psych verbs:
(1) The thunder frightened the baby.
Theme
Experiencer
(2) The children frightened the baby.
Theme
Experiencer
(3) Ali talked to Ayse.
Agent
Theme
Animacy is also an important quality that affects the degree of activeness but
the point of discussion should be whether to name the argument other than the Experiencer
as Theme in dyadic psych verbs or not.
In (2) the nature of the Theme argument is different from the Theme argument
of (3). In (2) the Theme is semantically the cause of the fear of the baby. Therefore this
type of psych verbs are regarded as different from the Experiencer Subject type psych
verbs in having a Causer argument.
Pesetsky (1996) decomposes the Theme role into two as the Target of Emotion
and the Subject Matter which is an essential distinction to differentiate between the Theme
arguments of the Experiencer subject and the Experiencer object verbs.
88
In Turkish, the Theme argument, parallel with the Experiencer argument is
either in the subject or in the object position. The surface case marking properties of the
Theme argument vary according to the structural position that it occupies; it is nominative
in the subject position and either accusative or oblique case marked in the object position.
II. 1. 3.
The Input of Psych Verbalization
This section gives a brief morphological analysis of psych verbs in Turkish.
Morphologically, psych verbs in Turkish are of three types:
1. non-derived simple base psych verbs
2. derived psych verbs
a. from nouns
b. from verbs
c. from adjectives
d. compound frozen forms
3. compound psych verbs
The majority of psych verbs are non- derived simple base verbs (the
etymological analysis of each of which is not the concern here) such as avun-, az-, bzk-,
bil-,
co~-, dii~iin-,
igren-, imren-, inan-, kzskan-, kzz-, kork-,
~a~-,
iirk-, iiz- etc.
In the second group, there are derived psych verbs. The psych verbs derived
from nouns constitute the biggest group. The majority of them
are
derived
nouns with -/An. Most of the time these verbs are intransitive:
Psych Noun+ -/An= Psych Verb
PSYCH NOUN, ' .
ac1
gurur
-Ian
-Ian
acllan- "to grieve"
gururlan- "to be arrogant"
from psych
89
h1rs
hiddet
h\iziin
ilgi
keyif
kibir
zevk
-Ian
-len
-len
-len
-len
-len
-len
Table I.
h1rslan- "to get angry"
hiddetlen- "to get furious"
hiizi.inlen- "to become sad"
ilgilen- "to be interested"
keyiflen- "to get into a good mood"
kibirlen- "to become arrogant"
zevklen- "to take pleasure"
Psych Noun+ -!An Psych Verb
Some psych verbs are derived from nouns with -!A. However, this time the
input noun can be either a psych or a non- psych noun:
Psych/ Non-Psych Noun+ -lA= Psych Verb
PSYCH/NONPSYCH NOUN
arzu
biiyii
etki
ip
yabanc1
yarg1
Table 2.
-lA
PSYCH VERB
-la
-le
-le
-le
-la
-la
arzula- "to wish for"
biiyiile- "to charm"
etkile- "to influence"
iple- "to mind"
yabancila- "to fmd smth. strange"
yarg1la- "to judge"
Psych/ Non-Psych Noun+ -lA- Psych Verb
The third group is the psych verbs derived from verbs. The base verb is either a
psych verb as in anla-m-lan-dzr-,
co~-ku-lan-, kzz-z~-.
san-gz-la-,
~a~-zr-
or not a psych verb
as in al-dzr-, bur-uk-Iah 9ek-in-, duy-um-sa-, kap-tzr-, oyna-t-.
Psych verbs are derived by five means from base verbs. The first way is first to
derive a psych noun (which mostly refers to a perception or at least implies it) with one of
the following morphemes and then to derive the psych verb from these nouns mostly either
with -lA or -!An:
Verb-> Psych Noun-> Psych Verb
INPUTV
al
bil
co~
duy
duy
es
kur
kur
san
-in9
-ku
-gu
-urn
-m
-gu
-untu
-g1
Table 3.
-len
-Ian
-Ian
-sa
-len
-la
-Ian
-la
bilin9len- "to become conscious"
co~kulan- "to get excited"
duygulan- "to be affected"
duyumsa- "to feel"
esinlen- "to be inspired"
kurgula- "imagine"
kuruntulan- "to worry for no reason"
sangila- "to suppose"
Verb-> Psych Noun-> Psych Verb
90
The second way is first to derive a psych adjective (all of which are the
modifiers of the Experiencer argument) mostly with -Gin and then to derive the psych
verb either with -iAn or -lA~:
Verb -> Psych Adjective -> Psych Verb
The third way to derive psych verbs from base verbs is to use the causative
morpheme -(D)Ir, or -t:
Verb+ Seemingly Causative Morpheme= Psych Verb
INPUTV
CAUSATIVE
\1\IQR.j)JiE.~
-d1r
al
kap
-tlr
-1r
-tir
kes
oyna
-t
-ir
-ur
u~
Table 5.
kapt1r- "to concentrate"
ka~1r- "to lose one's mind"
kestir- "to estimate"
oynat- "to become insane"
~i~ir- "to embarrass"
u~ur- "to lose one's mind"
Verb+ Causative Morpheme= Psych Verb
The fourth way to derive psych verbs from base verbs is to use the passive
morpheme -II deriving a reflexive sense:
Verb+ Seemingly Passive Morpheme= Psych Verb
91
Lastly, the following morphemes; reciprocal
-1~
and reflexive -In also derive
psych verbs from base verbs:
Verb+ -In/ -Ip Psych Verb
When we examine the sense relation between the base and the derived verb, the
question of whether the causative, passive, reflexive and reciprocal morphemes add their
original meaning to the input stem or they are just homophonous morphemes arises.
Further data is going to show that most of them are seemingly causativization or
passivization applications. Therefore, the mentioned suffixes can be regarded as
derivational suffixes without any voice effect.
In the fourth group there are psych verbs derived from adjectives which do not
have any other verbal counterparts, so, there is no other way of verbalizing them. Some of
these verbs are derived mostly from borrowed stative psych adjectives with -IA.r:
Psych Adjective+ -lA$= Psych Verb
PSYCH
ADJECTIVE
ahmak
ahk
aptal
-lA~
PSYCH VERB
-Ia~
ahmakl~-
-Ia~
ahkl~-
~Ilgm
-I~
-Ia~
hm;m
ifrit
sakin
salak
tedirgin
uysal
"to tum into a fool"
"to be astounded"
aptalla~- "to become stupid"
~IIgml~- "to become mad"
h1r~ml~- "to get ill tempered"
ifritle~- "to get angry"
sakinle~- "to get calm"
salakla~- "to become stupid"
tedirginle~- "to become uneasy"
uysalla~- "to become complaisant"
-Ia~
-le~
-le~
-Ia~
-le~
-Ia~
Table 8. Psych AdJeCtiVe+
lA~
- Psych Verb
92
Some of the verbs derived from adjectives are derived with the morpheme -!An
but the input adjective may not be a psych adjective as in the last one:
Psych/ Non-Psych Adjective+ -/An= Psych Verb
The difference between -!An and -lA~ verbs is that the former may occur with
an optional dative marked argument although the latter are truly intransitive (see Kural
(1996) for unaccusative vs. unergative distinction).
Some of these verbs are derived with -sA and the rest are the change of state
verbs derived with -Ar, -AI, or -An:
Psych/ Non-Psych Adjective+ -sA= Psych Verb
Non-Psych Adjective+ -Arl -AI= Psych Verb
Lastly there are some psych verbs which are derived from non native
compound bases. These verbs are frozen forms therefore the helping verb and the noun
component are merged into one verb form as in hazzet-, hisset-, kahrol-.
93
To conclude, there are 13 means for the derivation of psych verbs in Turkish
which we can summarize in the following formulas:
1. Non-Derived Base-> Psych Verb
2. Psych Noun+ -/An= Psych Verb
3. Psych/ Non-Psych Noun+ -/A= Psych Verb
4. Verb-> Psych Noun-> Psych Verb
5. Verb-> Psych Adjective-> Psych Verb
6. Verb+ Causative Morpheme= Psych Verb
7. Verb+ Passive Morpheme= Psych Verb
8. Verb+ ReciprocaV Reflexive Morpheme= Psych Verb
9. Psych Adjective+ -lA$= Psych Verb
10. Psych/ Non-Psych Adjective+ -!An= Psych Verb
11. Psych/ Non-Psych Adjective+ -sA= Psych Verb
12. Non-Psych Adjective+ -Ari-A!= Psych Verb
13. Compound Base -> Psych Verb
II. 1. 4.
Secondary Psych Senses
The analysis has shown that the majority of the psych verbs have psych senses
as their primary meaning which we can refer to as true psych verbs which include igren-,
imren-, ho$lan-, kork-, ozen-, sinirlen-,
§G§-,
tiksin-, yzl-.
94
Among the verbs in our data, there are no psych verbs that ca.n be secondarily
used in different senses other than their psych senses.
On the other hand, psych meanings of some verbs are their secondary
meanings. Semantically these verbs primarily belong to other verb classes but extensions
of meaning secondarily make them psych verbs. The following are some examples:
II. 2.
Psych Verb Types in Turkish
As discussed in previous chapters psych verbs are problematic for
linking/mapping theories proposed for different languages in the literature. The crosslinguistic tendencies show that the unexpected mapping of these verbs is universal. Such
differences in their syntactic and semantic behavior lead linguists define distinct classes of
psych verbs on different bases. Belletti& Rizzi (1988) focused on different surface
markings of thematic roles. Levin (1993) classified these verbs with respect to transitivity
95
alternations. Grimshaw (1990) and Tenny (1994) have used aspectual criteria in order to
explain the different syntactic behavior that these verbs display.
With the aim of providing a partial description related to their semantic and
syntactic properties, first, we will propose a classification for Turkish psych verbs. Our
claim is that there are four types of psych verbs in Turkish when we consider the syntactic
position and the surface case markings of thematic arguments. The following
configurations are based commonly on accepted thematic roles the Experiencer and the
Theme.
In the first group there is an Experiencer argument in the subject position. As
expected, this argument is nominative case marked. In the object position there is a Theme
argument and again as expected, this argument is accusative case marked just like ordinary
Agent Patient verbs.
If we reverse the thematic roles of the first group, we get the second group.
These are the verbs whose subjects are nominative case marked Themes and objects are
accusative case marked Experiencers.
The third group consists of verbs which have a nominative case marked
Experiencer in the subject position and a Theme argument in the object position just like
the first group. The difference is that the Theme argument here is not accusative but
oblique case marked.
The verbs of the fourth group are different from the other three groups in that
they do not have a Theme argument but just a nominative case marked Experiencer
argument in the subject position.
From now on, we are going to refer to verbs of these four classes as Type 1,
Type 2, Type 3(a/ b) and Type 4 psych verbs in Turkish as the following table illustrates:
96
The following sections analyze each of these types according to their input of
verbalization, surface case marking and the syntactic distribution of their non sentential
arguments.
II. 2. 1.
Type 1 Psych Verbs
The verbs in this group correspond to commonly called fear typepsyqh verbs of
Grimshaw (1990) and Tenny (1994) in English, and temere class ofBelletti& Rizzi (1988)
in Italian. The mapping of the thematic roles and syntactic positions do not pose any
problems for linking theories since they differ in no way from ordinary transitive dyadic
predicates. The Experiencer is in the subject position and the Theme is in the object
position. The surface case marking of arguments also corresponds to what is expected from
a transitive dyadic predicate; a nominative subject and an accusative object:
(1) Ben bu konuyu anladtm.
Experiencer
(NOM)
Theme
(ACC)
I understood this subject.
(2) Ay§e bu i§i yok arzuluyor.
Experiencer Theme
(NOM)
(ACC)
Ay~e wishes for this job
very much.
97
(3) Ane Ali'yi begeniyor.
Experiencer
(NOM)
Ay~e likes
Theme
(ACC)
Ali.
(4) Ay$e Ali'yi istiyor.
Experiencer
Theme
(NOM)
(ACC)
Ay~e wants Ali.
(5) Ali kansm1 k1skamyor.
Experiencer
(NOM)
Theme
(ACC)
Ali envies his wife.
(6) bgretmen ogrencileri ki.iyi.imsi.iyor.
Experiencer
Theme
(NOM)
(ACC)
The teacher looks down on the students.
(7) Cocuk annesini ozli.iyor.
Experiencer
(NOM)
Theme
(ACC)
The child misses his mother.
(8) Kadm bu adam1 seviyor.
Experiencer
Theme
(NOM)
(ACC)
The woman loves this man.
(9) Ali ailesini umursam1yor.
Experiencer Theme
(NOM)
(ACC)
Ali does not care about his family.
(10)
Ya$hlar genvleri 90k yad1rg1yor.
Experiencer
Theme
(NOM)
(ACC)
Old people consider the young people strange.
The input for the Type 1 psych verbalization is of five types:
1. a non-derived/ simple base: begen-, belle-, bil-, dile-, iste- ...
2. a verb: an-zm-sa-, anla-m-lan-dzr-, kes-tir-, kur-gu-la-, um-ur-sa- ...
3. a noun: arzu-la-, hesap-la-, ip-le-, onem-se--, yargz-la- ...
4. an adjective: agzr-sa-, garip-se-, hafifse- ...
5. a compound base: aff-et-, var-say-, zan-net- ...
98
A Type 1 psych verb can semantically be either mental as in algzla-, ammsa-,
kap-, kavra-, ogren- or emotional as in begen, kzskan, ozle, sev, yadzrga.
II. 2. 2.
Type 2 Psych Verbs
These verbs which correspond to frighten type psych verbs of Grimshaw (1990)
and Tenny (1994) in English, and preoccupare class of Belletti& Rizzi (1988) in Italian
are the problematic group for the linking theories since they map the Experiencer role to
the object position and the Theme role to the subject position. Therefore the Theme is
nominative and the Experiencer is accusative marked. That Type 2 psych verbs in Turkish
map their Experiencer and Theme roles in the same way with the other analyzed languages
shows this unexpected behavior to be a universal one. Moreover, in Turkish, (as it is the
case with most of the other languages) these verbs are limited in number. Some of the
verbs in the following examples require a form of support to emphasize the psych sense
such as a degree adverbial (9ok, fena etc.) or an instrumental PP. These verbs are either
secondary psych uses such as bog-, boz-, biiyiile-, kzr-, sars- as in p), (4), (5), (8), (9) or a
non native compound cezbet- as in (6) but not true lexical psych verbs like iiz- in {10). This
shows that most of the Type 2 verbs have somewhat a derived psych sense.
(1) Ali Ayseyi ayartti.
Theme Experiencer
(NOM) (ACC)
Ali seduced Ay~e.
(2) Konu§mact dinleyicileri bayd1.
Theme
(NOM)
Experiencer
(ACC)
The speaker bored the audience.
(3) Ane Ali'yi ticy ytldu ktskancyhgt ile bogdu.
Theme Experiencer
(NOM) (ACC)
Ane made Ali fed up with her jealousy for three years.
99
(4) Ali Ay~e'yi fena bozdu.
Theme Experiencer
(NOM) (ACC)
Ali embarrassed Ay~e badly.
(5) Kadm adarm tam~t1klan ilk saniyede biiyiiledi.
Theme Experiencer
(NOM) (ACC)
The woman charmed the man the ftrst second that they met.
(6)
Ay~e
Ali'yi sakinligi ile cezbetti.
Theme Experiencer
(NOM) (ACC)
Ay~e
(7)
charmed Ali with her calmness.
Ay~e
Ali 'yi 90k etkiledi.
Theme Experiencer
(NOM) (ACC)
Ay~e
influenced Ali very much.
(8) Babam beni 90k k1rd1.
Theme
(NOM)
Experienccr
(ACC)
My father hurt me very much.
(9) Ay§e Ali'yi gidi~iyle sarst1.
Theme Experiencer
(NOM) (ACC)
Ay~e
(10)
gave Ali a shock with her going.
Ay~e
Ali'yi uzdu.
Theme Experiencer
(NOM) (ACC)
Ay~e
made Ali upset.
According to their input of verbalization, similar to Type 1, there are five
sources for this process:
1. a non-derived/ simple base: boz-, kzr-, sars-, iiz- ...
2. a verb: ayar-t-,
kz~kzr-t- ...
3. a noun: biiyii-le-, etki-le- ...
4. an adjective: ergin-/e-.
5.
a compound base: cezb-et-.
These verbs too, can be either mental as in aydmlat-, sma- or emotional as in
bay-, bog-, boz-, biiyiile-, etkile-.
100
II. -2~3~ - _ Type 3 Psych Verbs
...
.
'
These are 'intransitive verbs. Similar to Type 1 verbs they have an Experiencer
argument as the subjeCt bu~ they have a Theme argument as the indirect object. What is
different is the oblique case marki~g of the Theme argument. We can regroup the verbs of
Type 3 with respect to their case marking in the following way:
1. Type 3a: For Type 3 verbs with dative marked Themes which are the
greatest in number.
2. Type 3b: For Type 3 verbs with ablative marked Themes which are less
great in number than Type 3 a.
II. 2. 3. 1. Type 3a Psych Verbs
These verbs have dative marked Themes. The dative Theme semantically is a
kind of goal or cause of the mentioned mental or emotional state. These verbs optionally
allow a source or a secondary cause argument:
(1) Ane Ali'ye ac1d1.
Experiencer Theme
(NOM)
(OAT)
Ay~e
felt pity for Ali.
(2) Ay§e Ali 'ye danld1.
Experiencer Theme
(NOM)
(OAT)
Ane put out with Ali.
(3) Ali Ay!?e'ye yok giiveniyor.
Experiencer Theme
(NOM)
(OAT)
Ali trusts Ay~e very
much.
(4) Ay§e arkada~ma imreniyor.
Experiencer
(NOM)
Ay~e
Theme
(OAT)
longs for her friend.
101
(5)
Ay~e
Experiencer
(NOM)
Ay~e
kadere inamyor.
Theme
(OAT)
believes the faith.
(6) Ali yalamm1za kand1.
Experiencer
(NOM)
Theme
(OAT)
Ali was fooled with our lie.
(7) Ali bize kmld1.
Experiencer Theme
(NOM) (OAT)
Ali was offended at us.
(8) Ali bize k1zd1.
Experiencer Theme
(NOM) (OAT)
Ali got angry with us.
(9) Ali
Ay~e'ye
kiistii.
Experiencer Theme
(NOM)
(OAT)
Ali put out with
(10)
Ay~e.
Genc;ler ~imdi Tarkan gibi pop ylld1zlanna ozeniyor.
Experiencer
(NOM)
Theme
(OAT)
Young people consider popstars like Tarkan as a model nowadays.
Mostly, the verbs of Type 3a are emotional as in begen-, giicen-, heveslen-,
inan-, kzrzl-, tutu!-, vurul- however; there are mental verbs too such as aldan-, kan-,
~a$zr-.
When we look at the inputs of the verbalization process we see that the five
sources for the other two groups are valid for Type 3a verbs too:
1. a non-derived/ simple base: dal-, imren-, inan-, kan-, kzz-, kiis-, ozen-,
yiiksiin- ...
2. a verb: ai-m-, bay-zl-, boz-ul-, kap-tzr-, /ar-zl-, tut-u!- ...
3. a noun: cezbe-len-, dert-len-, ejkar-lan-, hiiziin-len-, odak-lan-, ofke-len-,
iimit-len- ...
4. an adjective: aksi-len-, deli-len-, dik-len-, ters-len-, ye$il-len-, zor-sun- ...
5. a compound base: kahr-ol- and sabr-et-.
102
II. 2. 3. 2. Type 3b Psych Verbs
These verbs have ablative marked Themes which may be iiitetptefed as lhe
source or the cause of the mental or emotional state.
( 1) Adam hakaretlerden bezmi~ti.
The~
Experiencer
(NOM)
(ABL)
The man was fed up with the insults.
(2) Cocuklar sebze yemeginden b1kt1.
Experiencer
(NOM)
Theme
(ABL)
The children were fed up with vegetable meals.
(3) Ali kalabahktan yekindi.
Experiencer
(NOM)
Theme
(ABL)
Ali was ashamed of the crowd.
(4) Ane Ali ' den ho~lamyor.
Experiencer
(NOM)
Ay~e
Theme
(ABL)
likes Ali.
(5) Herkes bu sualar depremden korkuyor.
Experiencer
(NOM)
Theme
(ABL)
Everyone fears the earthquake nowadays.
(6) Polis adamm kansmdan ku~kulamyor.
The~
Experiencer
(NOM)
(ABL)
The police are suspected of the man's wife.
(7) Ogrenciler bu dersten sogumu~tu.
Experiencer
(NOM)
Theme
(ABL)
The students have lost their interest in this course.
(8) Ali bahktan tiksinir.
Experiencer
(NOM)
Theme
(ABL)
Ali feels disgust for the fish.
(9) Adam kimseden utanm1yor.
Experiencer
(NOM)
Theme
(ABL)
The man is not ashamed of anyone.
(10)
Ay$e riizgar sesinden bile urkuyordu.
Experiencer
(NOM)
Theme
(ABL)
Ane was scared even of the sound of the wind.
103
Mostly these verbs are emotional such as bzk-, bunal-, ho~lan-, igren-, soguthough they may be mental in a lesser number such as ku~kulan-,
~iiphelen-.
These verbs allow only animate beings to be their subjects as m bunal-,
kaygzlan-, utan-, iirk-, yzl- but not inanimate ones.
The mentioned five inputs are also valid for this class:
1. a non-derived/ simple base: bzk-, bunal-, kork-, tiksin-, usan- ...
2. a verb: t;ek-in-, es-in-len-, szk-zl-, um-ut-lan- ...
3. a noun: huy-lan-, kaygz-lan-,
4. an adjective: dar-a/-,
ku~ku-lan-, ~iiphe-len-
ho~-lan-, ho~-la~-
...
...
5. a compound base: haz-z-et-.
Nilsson ( 1985) uses the term predicates of "mental reaction" for the psych
verbs and analyzes them as two groups; one requiring the ablative and the other allative
(dative) case. In her analysis, the first group consists of verbs like bzk-, usan-, sogu-, nefret
et- and the second group consists of verbs like bayzl-, day- etc.
She claims that verbs of mental reaction which take the allative case like dan/-,
giiven-, inan- have an object which can be qualified as an abstract goal. These verbs
describe an attitude ofX towards Y.
The other group which take the ablative like duygulan-, umutlan-, yoru/describe a certain reaction of X due to Y and therefore, the object can be qualified as an
abstract source or a point of reference.
Nilsson puts forward that there are some other verbs which allow both the
allative and the ablative cases to be their objects including dertlen-, emin ol-, giicen-, incin-
, kaygzlan-, memnun ol-, razz of-, utan-.
104
Nilsson observes that this choice between the two cases is sensitive to the
temporal properties. For her (1) expresses a single future event whereas {2) expresses an
enduring certainty (1985:115):
(1) Derhal dedigimi yapacagma eminim.
(OAT)
I am sure that he will immediately do as I say.
(2) Derhal toparlanacag1mdan eminim.
(ABL)
I was sure that I would immediately pull myself together.
In such cases where the ablative and the allative cases seem to be the
alternatives of each other, Nilsson observes the following differences (1985: 116):
1. Ablative has a wider field of application than the allative (as (3) shows).
2. Ablative relates to continuous events or repeated instances
3. Ablative relates to events whose time component is quite indefinite,
therefore, there is a less specified! more diversified relation.
4. Allative refers to a more specified relation.
5. Allative notes afuture action (as (4) and (5) show).
6. Ablative notes either a factual or a potential action (as (6) and (7) show).
(3) Ben arkada~1mdan eminim.
(ABL)
I feel certain about my friend.
*Ben arkada~1ma eminim.
(OAT)
(4) Yanm1za girmeye utan1yor.
(OAT)
He is bashful to come up to you.
(5) Sokaga c1kmaya korkuyor.
(OAT)
He is afraid to go out.
(6) Utamyorum istemekten ama ...
{ABL)
It embarrasses me to ask for but ...
(7) Gelmeyi
di.i~i.inmekten
(ABL)
bile korkuyordu.
105
Even the thought of coming scared him.
This alternative choice between the two case markings of the same verb is not
just a feature of mental predicates. Nilsson gives examples from other groups of verbs
which allow for the same choice (1985: 117).
(8) a. Karde~im izmir' de yerle~ti.
(ABL)
b. Karde~im izmir'e yerle~ti.
(DAT)
Nilsson (1985) sees this freedom in the choice of case marking as a language
specific phenomenon and argues that this may be independent of lexical meaning but it is
rather related to the interpretation of the context and the "speaker's perception of the
relationship between the entities" (117).
In our analysis of Type 3a and 3b psych verbs, we have seen that both of these
case markings may imply a kind of cause ofthe mentaV emotional state of the Experiencer.
This is the reason why we added the notion of cause as a meaning component of both the
dative and the ablative marked arguments.
According to our analysis, in fact, the following examples may entail different
semantic interpretations with ablative and dative:
(9) Ali yurtd1~ma gitmeye korkuyor.
(NOM)
(1 0)
(DAT)
Ali yurtd1~ma gitmekten korkuyor.
(NOM)
(ABL)
In (10) Ali has gone to abroad before, has experienced something there and his
fear is related to the result ofthe activity that he already knows. However, in (9), this is the
first time that he will go to abroad and his fear is not related to the result (which is
unknown at that moment) but related to the event itself which may have potential (guessed,
expected or heard) unwanted results to be experienced.
106
There are two more intransitive psych verbs but they are not grammatically
significant to form a type on their own:
There is only one verb in our data that requires for a locative marked Theme
argument; yam!-. This is a mental verb whose input is a non-derived/ simple base:
(1) Ay$e bu konuda 90k yanthyor.
Experiencer
(NOM)
Theme
(LOC)
Ane was mistaken about this point.
There are seven verbs which have instrumental case marked Themes. All of
these verbs are semantically emotional type and all require for animate subjects.
(1) Ay$e eski hatualarla avunuyor.
Experiencer
Theme
(NOM)
(INST)
Ay~e consoles herself with
the old memories.
(2) Adam oglunun ba$ar1Slyla gururlantyordu.
Experiencer
(NOM)
Theme
(INST)
The man was proud of the success of his son.
(3) Ayse Ali ile ilgileniyor.
Experiencer Theme
(NOM)
(INST)
Ay~e is attracted
to Ali.
(4) Ktz gUzelligi ile kibirleniyordu.
Experiencer
(NOM)
Theme
(INST)
The girl was arrogant of her beauty.
One of the verbs in this class is derived from a simple base ; avun-, one of them
is derived from a verb;
~i~(in)-,
four verbs are derived from nouns; alaka(lan)-, gurur(lan)-
, ilgi(len)-, kibir(len)- and only one verb is derived from an adjective;
~en(len)-.
107
II. 2. 4.
Type 4 Psych Verbs
These verbs are distinct from the other intransitive verbs of Type 3 since they
do not have a Theme argument. Therefore the only argument that they have is a nominative
Experiencer which are mostly expressed in passive in English. This group of psych verbs
does not correspond to any of the defined psych verb classes in other languages. Therefore
it seems to exist only in Turkish.
(1) Ay~e afallad1.
Experiencer
(NOM)
Ay~e was bewildered.
(2) Kalabahk co~tu.
Experiencer
(NOM)
The crowd cheered.
(3)
Ay~e
son gtinlerde dalgtla~tL
Experiencer
(NOM)
Ay~e became
lost in thoughts recently.
(4) Dedem ya~land1k9a duruldu.
Experiencer
(NOM)
My grandfather calmed down as he got old.
(5) Ben bu ortamda 90k geri·liyorum.
Experiencer
(NOM)
I am tensed up very much in this atmosphere.
(6) Bebek huysuzl~t1.
Experiencer
(NOM)
The baby got uncomfortable.
(7) Baham 9abuk parlar.
Experiencer
(NOM)
My father gets furious quickly.
(8) Bu ktz iyice u9urdu.
Experiencer
(NOM)
This girl really went insane.
108
(9) Kadm bir ttirlii yatt~mtyordu.
Experiencer
(NOM)
The woman could not be calmed down in any way.
Annem dfule oranla biraz yumu~adt.
(1 0)
Experiencer
(NOM)
My mother became a little more tolerant compared to yesterday.
Only a few number of these verbs are mental such as
salakla~-
as well as
alzkla~-.
are emotional such as
ahmakla~-. aptalla~-.
bilin9len-, tozut-, u9-. u9ur- etc. The rest of the Type 4 verbs
co~ku/an-, co~-.
kudur-, par/a-,
uysalla~-. iirkek/e~-
etc.
The four inputs for the verbalization are as follows:
1. a non-derived base: afalla-,
apz~-.
az-, bat-, bit-,
co~-.
90k-, dol-, sin ...
2. a verb: These verbs are mostly derived with -/An or -/A~ either from nouns
such as
co~-ku-lan-,
bil-in9-len- or from adjectives such as
~a~-km-la~-. dur-gun-la~-.
kasz-n/ /-, siiz-iil-,
yat-z~-
co~-kun-la~-.
There are also other verbs like dur-ul-, ger-il-,
which directly derive from verbs with seemingly
voice suffixes.
3. a noun: These verbs are derived from nouns with -/An such as akli-lan-,
dioi-len-, gam-ian-, htrs-lan-, his-len- ...
4. an adjective: These verbs are derived from adjectives with
ahmak-laoi-. ftlgm-la§-, hzr9m-/a§-, sakin-le§-,
tedirgin-le~-
-/A~
such as
...
As the above analysis shows there are four basic types of psych verbs with
similar derivational patterns in Turkish. The first and the second type are transitive and the
others are intransitive; one with a dative, one with an ablative and one with a null Theme
argument.
109
II. 3.
Voice and Psych Verbs in Turkish
Although there are some common logical constraints for voice, there are no
universal structures which are uniform among different languages.
Voice in Turkish changes the surface case marking of arguments. Traditional
grammar books define voice as a property of verbs with respect to their subjects and
objects. Thus, voice is analyzed from two different points of view.
According to their subjects there are four types of voices and four types of
verbs which belong to these types; active, passive, reflexive and reciprocal. According to
their objects, verbs are of two types which are transitives and intransitives:
.
VOICE IN TURKISH
According to the properties of the subject
According to the properti~s of ~Jle objec~ .
Transitive
Intransitive
Active
Passive
Reflexive
Reciprocal
Table I.
Votce m Turktsh
Voice suffixes in Turkish can also function as derivational suffixes and this
sometimes causes confusions. The very often .confusion is among the passive and the
reflexive voice markers. This is sometimes regarded as a result of the fact that one suffix
has the mentioned both functions in the old Turkish. In some cases only the phonological
conditions determine the attachment ofthe suffixes.
Tietze (1989) argues that there is an ongoing merge of passive and the reflexive
forms in Turkish i.e. some verbs which have the reflexive marking have a passive meaning
and vice versa. Therefore overt morphological marking does not always correspond to the
semantic functions. His basic claim is that all of the verbs with the morpheme -11 were
passive in their origin and some of them acquired a reflexive meaning thereafter. The
following psych verbs from his list have the passive morpheme -11 but a "medioreflexive"
110
meaning (1989:287). Nakipoglu (1988) regards them as reflexives. These verbs do not
have -(I)n forms; bogul-, bozul-, ezil-, siiziil-, iiziil-, vurul-, yzkzl-..
In the other group in his list, there are verbs possessing the two forms (some of
which have psych meanings):
II. 3. 1.
Psych Verbs and Transitivity in Turkish
According to Taylan (nd.), Turkish psych verbs constitute a separate group of
intransitive verbs. She claims that psychological verbs are the only intransitive verbs which
take complements:
Transitive
v~rbs
+
+
Subject
Direct object complement
Other type of complement
Adjunct
+
Table I.
,.,
>
Intransitive verbs
..
'. ,, ·'
-~
",,
Psychological verbs
+
Other Intransitive verbs
+
-
-
+
+
..
+
Psych Verbs and Other lntraosrtlves
In our analysis, of the four types of psych verbs in Turkish, only the Type 1 and
Type 2 are transitive and the rest are intransitive:
TYPE .
1
2
+
+
+
+
+
3a
3b
4
Table 2.
Transitivity in Turkish Psych Verbs
111
It is possible in Turkish optionally to delete the arguments of Type 3a and 43b
whenever the discoursal factors allow. The following verbs of Type 3a for example allow
the deletion of their dative marked Themes requiring mostly (but not always) an adverbial
support (iyice, fena etc.):
These verbs of Type 3a are most of the time derived with either -/An or -/A~:
TYPE
NOUN/
ADJECTIVE+-
With:(DAT)
()p#()nal.!).~letiqn
Ay~e
Ay~e duygulandt.
(Ayse was affected)
of (DAT)
(lAn)/-(lA~)
3a
duy-gu-lan-
filme duygulandt.
(Ayse was affected by the film.)
3a
3a
3a
huy-suz-lansev-da-lanaksi-len-
Bebek gUrilltilye huysuzlandt.
Bebek huysuzlandt.
(The baby became uncomfortable because of the
noise.)
Ali Ay~e'ye sevdalandt.
(Ali felt in love with Ane.)
(The baby became uncomfortable.)
Dede torunlanna aksilendi.
(The grandfather
grandchildren.)
3a
3a
3a
3a
3a
deli-lendik-le-
became
obstinate
Ali sevdalandt.
(Ali felt in love.)
Dede aksilendi.
to
his
(The grandfather became obstinate.)
Kocas1 bu sozlere delilendi.
Kocas1 delilendi.
(Her husband went insane for these words.)
(Her husband went insane.)
Adam patronuna diklendi.
Adam diklendi.
(The man got stubborn to his boss.)
(The man got stubborn.)
heves-len-
<;ocuklar tatile heveslendi.
<;ocuklar heveslendi.
(The children became desirous.)
hiiziin-len-
(The children desired for the holiday.)
Ay~e bu ~ark1lara hUzOnleniyor.
(Ane becomes sad because of these songs.)
Ay~e
Annem mutlulugumuza keyiflendi.
Annem keyiflendi.
keyif-len-
3a
ters-len-
3a
duy-ar-siz-la~-
hilzOnleniyor.
(Ane becomes sad.)
(My mother took pleasure of our happiness.)
(My mother took pleasure.)
Ali babasma terslendi.
Ali terslendi.
(Ali got befouled to his father.)
(Ali got befouled.)
<;ocuk hakaretlere duyars~zl~tt.
<;ocuk duyars~zl~tJ.
(The child was desensitized with the insults.)
(The child was desensitized.)
.
.
Table 4. Opttonal Argument Deletion of Denved Type 3a Verbs
There are also verbs of Type 3b which allow the optional deletion of their
ablative Themes:
TYPE VERB
3b
endi~e-len-
Ailesi hastahgmdan
endi~eleniyor.
(Her family worries about her illness.)
3b
gocun-
Ay~e
sozlerimden gocundu.
(Ane took offense at my words.)
endi~eleniyor.
(Her family worries.)
Ailesi
Ay~e
gocundu.
(Ane took offence.)
112
3b
3b
huy-lankaygt-lan-
3b
onur-lan-
3b
sogu-
II. 3. 2.
Ali adamm tavtrlanndan huylandt.
Ali huyland1.
(Ali got uneasy with the man's attitudes.)
(Ali got uneasy.)
Ailesi hastahgmdan kaygilamyor.
Ailesi kaygtlamyor.
(Her family worries about her illness.)
(Her family worries.)
Ali ~ok onurlandt.
davrant~tan ~ok onurlandt.
(Ali felt honored very much with this behavior.)
(Ali felt honored very much.)
Arkad~lan ondan sogudu.
Arkad~lan sogudu.
(His friends lost their love for him.)
(His friends lost their love.)
Table 5. Opttonal Argument Deletion of Type 3b Verbs
Ali bu
Psych Verbs and Causativity in Turkish
Causativity can be regarded as a transitivizing operation which makes
intransitive verbs transitive adding a Causer argument to the argument structure. If the verb
is already
a transitive verb, the causativity has the function of increasing the degree of the
transitivity ofthe verb.
There are two types of causativity in Turkish; morphological and periphrastic.
Periphrastic causatives are formed with predications like sag/a-, neden ol- etc.
Morphological causatives on the other hand uses one of the causative suffixes -(D)lr, and(l)t. The affix to be attached is determined by the phonological conditions.
Multiple causation is also possible in Turkish. Theoretically there is no limit to
the addition of the Causer arguments but, as the number of the Causers and Causees
increase, it gets harder to interpret the causative events semantically. Thus, there is a
semantic restriction (though not logical) on the number of the Causers to be added, not just
for psych causatives but for all causative events. Therefore, the addition of Causer
arguments is limited with the causative events which can be interpretable. It may also have
discoursal and stylistic functions but beyond the interpretability boundary, multiple
causation is a vacuous application.
113
This section analyzes the relationship between the psych verb types in Turkish
and causativity. The analysis shows that morphologically there are five causativity types
valid for all psych verb classes:
1.
Morphological Causatives: They utilize one of the phonologically determined
causative morphemes.
1. 1.
There are psych verbs that can be causativized which we will refer to as(+)
1. 2.
There are psych verbs only the causative forms of which have psych
Caus.
meanings which we will refer to as Only Caus.
1. 3.
There are verbs that can be causativized but the causative form and the non-
causative form of the verb have different meanings which are both psych which we will
refer to as Diff Psych.
1. 4.
There are psych verbs only the non-causative forms of which have psych
meanings which we will refer to as Only Non Caus.
1. 5.
There are psych verbs that can not be causativized which we will refer to as
(-) Caus.
2.
ol-.
Periphrastic causatives: They utilize one of the predications like sag/a- and neden
114
II. 3. 2. 1.
Morphologically Causative Psych Verbs
II. 3. 2. 1. 1.
Type 1 Psych Verbs and Morphological Causativity
Type 1 has psych verbs which belong to five of the mentioned classes:
(+)CAUS •..
affet -(Dir)ammsa-(t)ba~t~la -{t)be~en -{Dir)benimse -(t)dti~Un -{Dir)hattrla -(t)hisset -(Dir)kavra -{t)ktskan -(Dir)sev -(Dir)sez -(Dir)unut -{Dir)-
.
anlamlan -(Dlr)oyna -{t)ye -(Dir)bat -(lr)c;tk -{lr)-
c;ak:t c;ak -(Dir)kap:t kap -(Dir)-
Table L
anla -{t)iste -{t)ktvlr -(t)-
alglla -{t)arzula -{t)dile -(t)garipse -(t)kotiimse -{t)kurgula -{t)onemse -(t)ongor -{Dir)san -{Dir)um -(Dir)varsay -(Dir)yarglla -{t)zannet -{Dir)-
Type I Psych Verbs and Causativity
In the first class there are psych verbs which can be causativized such as:
(1) a. Ali
Ay~e'yi
E~tperiencer
(NOM)
Ali forgave
b. Ali
affetti.
Theme
(ACC)
Ay~e.
Ay~eyi
Causer Theme
(NOM) (ACC)
Ahmet'e affettirdi.
Experiencer
(OAT)
Ali made Ahmet forgive
Ay~e.
c. Ali Ahmet'e Ay§e'yi affettirdi.
Causer Experiencer
(NOM)
(OAT)
Theme
(ACC)
Ali made Ahmet forgive
Ay~e.
d. Ali Ay§e'yi affettirdi.
Causer
(NOM)
Theme
(ACC)
Ali made
Ay~e
be forgiven.
e. *Ali Ahmet' e affettirdi.
Causer Experiencer
(NOM) (OAT)
*Ali made Ahrnet forgive.
f. Ali kendini Ay§e'ye affettirdi.
Causer
(NOM)
Theme
(ACC)
Experiencer
(OAT)
Ali made himself be forgiven by
Ay~e.
115
g. Ali Ay~e'ye kendini affettirdi.
Causer Experiencer
(NOM)
(DA T)
Theme
(ACC)
Ali made himself be forgiven by Ane.
h. Ali kendini affettirdi.
Causer
Theme
(NOM) (ACC)
Ali made himself be forgiven.
(2) a. Ali bu fikri benimsedi.
Experiencer Theme
(NOM)
(ACC)
Ali adopted this idea.
b. Ali Ay§e'ye bu fikri benimsetti.
Causer
Experiencer
(NOM)
(DA T)
Theme
(ACC)
Ali made Ane adopt this idea.
(3)
a.~
Turk kahvesini sevdi.
Experiencer
Theme
(NOM)
Ay~e loved
(ACC)
the Turkish coffee.
b. Ahmet Ay$e'ye Turk kahvesini sevdirdi.
Causer
Experiencer
(DAT)
made Ay~e
Theme
(NOM)
Ahmet
(ACC)
love the Turkish coffee.
c. Ali Ay$e'yi sevdi.
Experiencer Theme
(NOM)
(ACC)
Ali loved Ay~e.
d.? Ahmet Ali'ye Ay§e'yi sevdirdi.
Causer Experiencer Theme
(NOM)
(DA T)
Ahmet made Ali love
(ACC)
Ay~e.
(4) a. Ali kotu haberi sezdi.
Experiencer
Theme
(NOM)
(ACC)
Ali felt the bad news.
b. Ali Ay$e'ye koru haberi sezdirdi.
Causer Experiencer
(NOM)
(DA T)
Theme
(ACC)
Ali implied the bad news to Ay~e.
(5) a. Ali eski gunleri unuttu.
Experiencer
Theme
(NOM)
(ACC)
Ali has forgotten the past.
b. Ali
Ay~e'ye
Causer Experiencer
(NOM)
(DA T)
eski gtinleri unutturdu.
Theme
(ACC)
Ali made Ane forget the past.
116
In (la), the verb affet- has a nominative subject and an accusative object. In the
causative form (1 b), a Causer argument is added to the subject position, the Theme is still
accusative but the Experiencer is dative.
( 1c) shows that the reversal order of the Theme and the Experiencer is also
possible keeping the case marking of the thematic arguments the same. Whichever
argument is in the preverbal position, it is the focused one. In (ld), we see that the Causer
argument and the accusative marked Theme argument can be used without the
Experiencer. This accusative Theme can also be a reflexive pronoun with or without the
dative Experiencer as in (1 f) and (I g). However, the Causer and the Experiencer without
the Theme is ungrammatical as in (le). The English counterpart missing a Theme
argument is also ungrammatical. All these can be formulated as:
'J:'yp~
Typel Psych Construction
Experiencer + Theme
(NOM)
t ~~M~~tiY~,,~,~y~Jt.{;;~tl§t~gcti9~ .
'
Causer+ Experiencer+ Theme
(ACC)
(NOM)
(OAT)
(ACC)
Causer+ Theme+ Experiencer
(NOM)
(ACC)
{OAT)
Causer+ Theme
(NOM)
(ACC)
* Causer+ Experiencer
(NOM)
Table 2.
{OAT)
Type I Causative Psych Construction
(3b) also seems to be more acceptable than (3d). The verb sev1- in (3b) means
"changing one's attitude in a more positive way" but sev2- in (3c) and (d) means "to have
romantic feelings for someone or to feel attracted to someone".
However, there are some exceptions to the causativity rules above. In the
following sentences, there are psych verbs which have a nominative Causer and an
accusative Theme ((6)- (12) sentences). These sentences are normally interpreted as if
there is an implied Experiencer argument. This dative marked Experiencer is optional ((b)
examples).
117
(c) and (d) examples suggest the two alternative interpretations of (a)s. In (c)s,
there is a nominative Causer and a genitive Theme with a periphrastic causative which
nominalizes the verb. In (d)s, there is a nominative Causer and a genitive Theme with an
optional Experiencer. Here, similarly, there is a periphrastic causative nominalizing the
passivized verb. This kind of a passive interpretation can also be observed in the English
translations of the sentences.
From (6) to (12) of the following, all sentences normally accept the
interpretation (d) rather than (c):
(6)a. Ali Ayse'yi affettirdi.
Causer
Theme
(NOM)
(ACC)
Ay~e
Ali made
b. Ali
be forgiven.
Ay~e'yi
herkese affettirdi.
Experiencer
(OAT)
c. *Ali Ay~e'nin affetmesini sagladt/ neden oldu.
d. Ali
Ay~e'nin
(herkes tarafmdan) affed(il)mesini sagladt/ neden oldu.
(7)a. Ali Ayse'yi an1msatt1.
Causer
Theme
(NOM)
(ACC)
Ali made Ay~e be remembered.
b. Ali
Ay~e'yi
herkese ammsatt1.
Experiencer
(OAT)
c. *Ali
Ay~e'nin
arumsamasm1 sagladtl neden oldu.
d. Ali Ay~e'nin (herkes tarafmdan) antmsa(n)masmt sagladtl neden oldu.
(8)a. Ali Ane'yi bagt~lattt.
Causer
Theme
(NOM)
(ACC)
Ay~e be
Ali be
b. Ali
forgiven .
Ay~e'yi
herkese bagt~lattt.
Experiencer
(OAT)
c. *Ali
Ay~e'nin bag1~lamasm1
sagladt/ neden oldu.)
118
d. Ali Ane'nin (herkes tarafmdan) bagt~la(n)masmt sagladt/ neden oldu.
(9)a. Ali Ayse'yi begendirdi.
Causer
(NOM)
Theme
(ACC)
Ali made Ane be liked.
b. Ali
Ay~e'yi
herkese begendirdi.
Experiencer
(OAT)
c. *Ali Ay~e'nin begenmesini sagladt/ neden oldu.
d. Ali
Ay~e'nin
(herkes tarafmdan) begen(il)mesini sagladt/ neden oldu.
(lO)a. Ali Ayse'yi benimsetti.
Causer
(NOM)
Ali made
b. Ali
Theme
(ACC)
Ay~e be
Ay~e'yi
accepted.
herkese benimsetti.
Experiencer
(DAT)
c. *Ali
d. Ali
(II)
Ay~e'nin
Ay~e'nin
benimsemesini sagladt/ neden oldu.
(herkes tarafmdan) benimse(n)mesini sagladt/ neden oldu.
a. Ali Ayse'yi sevdirdi.
Causer
(NOM)
Ali
Theme
(ACC)
made Ay~e be
b. Ali
Ay~e'yi
loved.
herkese sevdirdi.
Experiencer
(DAT)
(12)
c. *Ali
Ay~e'nin
d. Ali
oldu.
Ay~e'nin
sevmesini sagladt/ neden oldu.
(herkes tarafmdan) sev(il)mesini saglad1/ neden
a. Ali Ayse'yi unutturdu.
Causer.
(NOM)
Theme
(ACC)
Ali made Ane be forgotten ,
b. Ali Ay~e'yi herkese unutturdu.
Experiencer
(DAT)
c. *Ali Ay~e'nin unutmasm1 saglad1/ neden oldu.
119
d. Ali Ane'nin (herkes tarafmdan) unut(ul)masmt sagladt/ neden
oldu.
However, in (13a) and (14a)
Ay~e
is an accusative Experiencer not a Theme
unlike (6)- (12) above. Therefore (b)s are unacceptable and the normal interpretations are
(c)s but not (d)s:
a. Ali Ayse'yi du~fuldiirdu.
(13)
Causer Experiencer
(NOM)
(ACC)
Ay~e
Ali made
think.
b. *Ali Ayse'yi herkese du~iindiirdu.
Causer Experiencer
(NOM)
Theme
(ACC)
c. Ali Ane'nin
(DA T)
dii~unmesini
sagladt/ neden oldu.
d. *Ali Ane'nin (herkes tarafmdan)
oldu.
(14)
dii~Un(Ul)mesini
sagladt/ neden
a. Ali Ayse'yi ktskandtrdt.
Causer Experiencer
(NOM)
(ACC)
Ali made Ay~e envy.
b. *Ali Ayse'yi herkese ktskandtrdt.
Causer Experiencer
(NOM)
c. Ali
Theme
(ACC)
Ay~e'nin
(DA T)
ktskanmasmt sagladt/ neden oldu.
d. *Ali Ay~e'nin (herkes tarafmdan) ktskan(tl)masmt sagladt/ neden
oldu.
The conclusion could be that the verbs in (13) and (14) are not the causative
forms of the verbs kzskan- and
du~un-
rather they are separate lexical entries which belong
to Type 2 psych verbs:
(15) a. Ali Ayse'yi ktskandt.
(Type 1)
Experiencer Theme
(NOM)
(ACC)
b. Ali Ayse'yi ktskandtrdt.
(Type 2)
Theme Experiencer
(NOM)
(ACC)
c. Ali Ayse'yi dii~Undii.
Experiencer
Theme
(NOM)
(ACC)
(Type 1)
120
d. Ali Ay§e'yi du~fuldurdu.
(Type 2)
Theme Experiencer
(NOM) (ACC)
For these verbs, the rule above can be restated adding the last two formulas:
Type 1 Psych Construction
Experiencer + Theme
(NOM)
Type 1 Causative Psych Construction
Causer+ Experiencer+ Theme
(NOM)
(ACC)
(OAT)
(ACC)
Causer + Theme+ Experiencer
(NOM)
(ACC)
(OAT)
Causer+ Theme
(NOM)
(ACC)
* Causer+ Experiencer
(NOM)
(OAT)
Causer+ Experiencer
(NOM)
(ACC)
*Causer+ Experiencer+ Theme
Table 3.
(NOM)
(ACC)
Type I Causative Psych Construction (Revised)
(OAT)
There are some verbs of Type 1 which can not be causativized:
(1)
a.~
Ali'yi arzuladt.
Experiencer Theme
(NOM)
(ACC)
Ay~e
desired for Ali.
b. *Ahmet Ay§e'ye Ali'yi arzulatt1.
Causer
(NOM)
Experiencer Theme
(OAT)
(ACC)
* Ahmet made Ay~e desire for Ali.
(2) a. Ay§e evliligi diledi.
Experiencer Theme
(NOM)
(ACC)
Ay~e
wished for marriage.
b. *Ali Ay§e'ye evliligi diletti.
Causer Experiencer Theme
(NOM) (OAT)
(ACC)
*Ali made Ay~e wish for marriage.
(3) a. Ali bu davrant§lart garipsedi.
Experiencer
(NOM)
Theme
(ACC)
Ali found these behaviors strange.
b. *Ay§e Ali 'ye bu davram§lart garipsetti.
Causer Experiencer
(NOM) (OAT)
Theme
(ACC)
* Ay~e made Ali find these behaviors strange.
(4) a. Ali bu dersi hafifsedi.
Experiencer
(NOM)
Theme
(ACC)
Ali took this course lightly.
121
b. *Arkadaslan Ali 'ye bu dersi hafifsetti.
Causer
(NOM)
Experiencer Theme
(DA T)
(ACC)
*His friends made Ali take this course lightly.
(5) a. Ayse ailesini her zaman onemsedi.
Experiencer
(NOM)
Ay~e
Theme
(ACC)
has always cared about his family.
b. *Ali Ayse'ye ailesini onemsetti.
Causer Experiencer Theme
(NOM) (DA T)
(ACC)
*Ali made Ay~e care about her family.
When the Causer is inanimate, causative forms of these verbs especially with a
generic Experiencer seem more acceptable (although they are not perfect):
(6) a. ?Bu reklam bana kola is;meyi arzulattyor.
Causer
(NOM)
Experiencer Theme
(DA T)
(ACC)
This commercial makes me desire for cola.
b. Bu reklam insana kola is;meyi arzulattyor.
Causer
(NOM)
Experiencer Theme
(DA T)
(ACC)
This commercial makes people desire for cola.
(7) ?Onlarm mutlulugu Ali 'ye evliligi diletti.
Causer
(NOM)
Experiencer
(DA T)
Theme
(ACC)
Their happiness caused Ali to wish for marriage.
(8) ?Gordtiklerim ve duydukl<:ll'Im bana da bu davran1slar1 garipsetti.
Causer
(NOM)
Experiencer
(DA T)
Theme
(ACC)
What I had seen and heard caused me too to find these behaviors strange.
(9) ?Smavda ald1g1 iyi notlar Ali'ye bu dersi hafifsetti.
Causer
(NOM)
Experiencer Theme
(DA T)
(ACC)
Good grades that he got made Ali take this course lightly.
(10) ?Ayse'nin telkinleri Ali'ye ailesini onemsetti.
Causer
(NOM)
Ay~e's
Experiencer Theme
(DAT)
(ACC)
consolations made Ali care about his own family.
122
II. 3. 2. 1. 2. Type 2 Psych Verbs and Morphological Causativity
As we mentioned in the previous chapters, there is a Ut1iversality in the
unexpected behavior of Type 2 psych verbs. Sometimes the aspectual properties were seen
as the reason for this. The most thoroughly discussed of these properties is the causation
which is assumed to be inherent in the meaning of these verbs.
Our analysis of Type 2 psych verbs in Turkish shows that none of these verbs
can be used causatively:
(+) CAUS.
DlFF. PSYC_H ···· · •· \ ; ~9N- Cj\l,JS.
0
0
ONLYCAUS.
ayar -(t)avu -(t)-kt~ktr -(t)-
0
Table 4.
. .'
~
J-)CAUS •
All of the Type 2 verbs
..
Type 2 Psych Verbs and Causat1v1ty
Therefore, the assumption that the semantic notion of causation is inherent in
the meaning of these verbs seems to be valid in Turkish too as the ungrammaticality of the
following causative counterparts demonstrate:
( 1) a.
Konu~mact
Theme
(NOM)
dinleyicileri baydt/ stktt.
Experiencer
(ACC)
The speaker bored the audience.
b. *Adam konu$mactya dinleyicileri baydtrdt/ stkttrdt.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*The man made the speaker bore the audience.
(2) a. Ane Ali'yi herkesin oniinde bozdu.
Theme Experiencer
(NOM) (ACC)
Ane humiliated Ali in front of everyone.
b. *Ahmet Ali 'ye Ane 'yi herkesin oniinde bozdurdu.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*Ahmet made Ali humiliate Ay~e in front of everyone.
Another use of boz- as a compound with a possessive psych noun like keyfinil
moralini etc. in causative is acceptable:
123
c.~
Ali'nin moralini bozdu.
Theme
(NOM)
Ay~e
Experiencer
(ACC)
depressed Ali.
d. Ahmet Ane'ye Ali'nin moralini bozdurdu.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
Ahmet made Ay~e depress Ali.
(3) a.
Ay~e
Ali'yi yok etkiledi.
Theme
(NOM)
Ay~e
Experiencer
(ACC)
affected Ali deeply.
b. *Ahmet Ay~e'ye Ali 'yi yok etkiletti.
Causer
(NOM)
Theme
(DAT)
Experiencer
(ACC)
*Ahmet made Ay~e affect Ali deeply.
(4)
a.~
Ali'yi k1rd1.
Theme Experiencer
(NOM) (ACC)
Ay~e
hurt Ali.
b. *Ahmet Ali'ye Ay~e'yi kudud1.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*Ahmet made Ali hurt Ay~e.
(5) a. Ane Ali'yi sarsti.
Theme Experiencer
(NOM) (ACC)
Ay~e
gave Ali a shock.
b. *Ahmet Ay~e'ye Ali 'yi sarstud1.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*Ahmet made Ay~e give Ali a shock.
(6) a.
Ay~e
Ali'yi iizdii.
Theme Experiencer
(NOM) (ACC)
Ay~e
depressed Ali.
b. *Ahmet Ali'ye Ay~e'yi iizdiirdii.
Causer
(NOM)
Theme
(DAT)
Experiencer
(ACC)
*Ahmet made Ali depress Ay~e.
124
In fact there is not either a phonological or a semantic restriction to causativize
these verbs. A verb like Uzdtir- is both phonologically and semantically possible (as the
possibility of a periphrastic causative like iiziilmesini sag/a- (which will be discussed in the
related chapter) demonstrates). Thus, the restriction seems to be related to only the
morphological causativity.
Most of the Type 2 verbs above have primarily a non psych/ physical sense.
When these are used in their first sense it is possible to causativize them; like szk- lszk-tzr-,
boz- I boz-dur-, kzr- I kzr-dzr- etc.
There are only 3 verbs of Type 2 whose bases seem to be inseparable from the
causative morpheme:
(1) a. *Ali Ay§e'yi ayard1.
Theme Experiencer
(NOM) (ACC)
b. *Ahmet Ali 'ye Ay§e 'yi ayartt1.
Causer
(NOM)
Theme
(DA T)
Ahmet made Ali
Expcriencer
(ACC)
seduce Ay~e.
c. Ali Ay§e'yi ayartt1.
Theme Experiencer
(NOM) (ACC)
Ali seduced Ay~e.
d. Ahmet Ay~e'nin ayarttlmasmt sagladt/ neden oldu.
(2) a. *Ali Ay§e'yi aydmladt.
Theme Experiencer
(NOM) (ACC)
b. *Ahmet Ali'ye Ay§e'yi aydmlatt1.
Causer
(NOM)
Theme
(OAT)
Ahmet made Ali
Experiencer
(ACC)
enlighten Ay~e.
c. Ali Ay§e'yi aydmlattt.
Theme Experiencer
(NOM) (ACC)
Ali enlightened Ay~e.
d. Ahrnet Ay~e'nin aydmlanmasmt sagladt/ neden oldu.
125
(3) a. Ali Ayse'yi
kt~kudt.
Theme Experiencer
(NOM) (ACC)
b.* Ahmet Ali'ye Ane'yi
Causer
(NOM)
Theme
(DA T)
kt~ktrttt.
Experiencer
(ACC)
Ahmet made Ali incite Ay~e.
c. Ali Ane'yi
kt~ktrttt.
Theme Experiencer
(NOM) (ACC)
Ali incited Ane.
d. Ahmet Ane'nin kt~kumasmt sagladt/ neden oldu.
In (1), (2) and (3) above the base verbs can not be used without the causative
morpheme as in ( 1a), (2a) and (3a). The normal causative construction with the addition of
a nominative Causer argument is also ungrammatical as in (lb), (2b) and (3b). The only
possibility is the usual Type 2 construction; a nominative Theme and an accusative
Experiencer ((lc), (2c) and (3c) with a causative semantic interpretation as in (ld), (2d)
and (3d).
The grammatical and ungrammatical examples lead to the following rule for
Type 2 causative psych constructions:
Type 2 Psych Construct!on
Theme + Experiencer
(ACC)
* Causer+ Theme+ Experiencer
(NOM)
(NOM)
(DA T)
(ACC)
Theme + Experiencer
Table 5.
(ACC)
(NOM)
Type 2 Causative Psych ConstructiOn
Therefore, the only possibility to express causativity is to use the Type 2 psych
verbs in a normal psych construction. Thus we argue that these lexical causatives do not
permit for further morphological causativization which means that all of the causativity
rules for Type 1 verbs are invalid for Type 2 verbs.
126
II. 3. 2. 1. 3. Type 3a Psych Verbs and Morphological Causativity
Type 3a psych verbs have also verbs belonging to five of the classes mentioned
above:
(+) CAUS.
c,:tldtr -(t)danl-(t)duygulan -(Dlr)efkarlan -(Dlr)giicen -(Dlr)heveslen -(Dlr)huysuzlan -(Dlr)inan -(Dir)ktz -(Dlr)kiis -(Dlr)ofkelen -(Dlr)ozen -(Dlr)sabtrstzlan -(Dlr)sinirlen -(Dlr)-
ONLYCAUS.
DIFF.PSYCH
NON-CAUS.
(-)CAUS.
ai-(Dlr)-
0
dal-(Dlr)kopiir -( t)patla -(t)tellen -(Dlr)-
giiven -(Dlr)ktr-(Dlr)-
Table 6.
~a~akal-(Dlr)-
takll-(Dlr)vurul -(Dlr)-
..
Type 3a Psych Verbs and Causattvtty
The causativization with Type 3a verbs is the reverse procedure of the
causativization with Type 1 verbs:
Ty~e
1 causativization:
la. Ali Ay$e'yi affetti.
Ty~e 3a
Experiencer Theme
(NOM)
(ACC)
Ali forgave Ay~e.
Experiencer Theme
(NOM)
(OAT)
Ane believed Ali.
2a. Ahmet Ay$eyi Ali 'ye affettirdi.
2b. Ahmet Ali 'yi Ay§e'ye inandtrdl.
causativization:
lb. Ali Ay$e'ye inand1.
Causer Theme
Experiencer
(NOM) (ACC)
(OAT)
Ahmet made Ali forgive Ay~.
Causer Experiencer Theme
(NOM)
(ACC)
(OAT)
Ahmet made Ali believe Ay~.
3a. Ahmet Ali'yeAne'yi affettirdi.
Causer Experiencer Theme
(NOM) (OAT)
(ACC)
Ahmet made Ali forgive Ane.
Table 7.
3b. Ahmet Ane'ye Ali'yi inandud1.
Causer
Theme
Experiencer
(NOM)
(OAT)
(ACC)
Ahmet made Ali believe Ane.
Type I and Type 3a Causative Psych Constructton (Companson)
Recall that the causativization of Type 1 verbs is simply the addition of a
nominative Causer argument and the dative marking of the Experiencer argument ((la) and
(2a)). If we exchange the order of the Experiencer and the Theme keeping the case
marking of the thematic roles the same, the difference is nothing more than the emphasis.
This is also valid for Type 3a causatives ((3a) and (3b)).
127
When we causativize Type 3a verbs, similarly a nominative Causer argument is
added to the subject position. The dative Theme argument does not change but the
Experiencer is accusative marked ((1b) and (2b)).
One difference between Type 1 and 3a in terms of causativization is that Type
1 verbs allow the optional deletion of the dative Experiencer argument (4a) but Type 3a
verbs do not (4b). Moreover, Type 1 verbs do not allow the deletion of accusative Theme
argument (5a) but Type 3a verbs do allow the deletion of the dative Theme (5b):
Type 1 causativizatiob
4a. Ahmet Ay~e'yi affettirdi.
,,
~'
Type 3a ~ausati"yizalion
4b. *Ahmet Ay~e'ye inand1rd1.
('
Causer
Theme
(NOM)
(OAT)
*Ahmet made Ay~e be believed,
Causer
Theme
(NOM) (ACC)
Ahmet made Ay~e be forgiven,
Sa. *Ahmet Ali'ye affettirdi.
Causer Experiencer
(NOM) (OAT)
*Ahmet made Ali forgive,
Table 8,
,,
5b. Ahmet Ali 'yi inand1rd1.
Causer Experiencer
(NOM) (ACC)
Ahmet made Ali believe,
Type I and Type 3a Causat1ve Psych Construction (Differences)
As expected, verbs of Type 1 do not allow the deletion of the accusative
argument in a causative structure since an accusative marked argument is what these verbs
necessitate in non causative structures. The unexpected point is that although a Type 3a
verb requires for a dative marked argument in a non causative structure, it is
ungrammatical with a dative argument in causative structure and requires for an accusative
argument. This shows that causativity is not sensitive to argument structure and that it
necessitates the existence of an accusative argument no matter what the verb takes in non
causative structure.
The following sentences (6) and (7) are the verbs of Type 3a that can be
causativized. (6c) and (7c) show that these verbs allow the deletion of dative Theme
argument:
(6) a. Ali Ay$e'ye danld1/ gticendi/ ktistti.
Experiencer Theme
(NOM)
(OAT)
Ali put out with Ay~e,
128
b. Ahmet Ali'yi Ane'ye danlttl/ gucendirdi/ kusrurdu.
Causer Experiencer Theme
(NOM) (ACC)
(DA T)
Ahmet made Ali.
c. Ali Ayse'yi danltti/ gucendirdi/ kustfudu.
Causer Experiencer
(NOM) (ACC)
(7) a. Ali Ayse'ye k1zdi/ otkelendi/ sinirlendi.
Experiencer Theme
(NOM)
(OAT)
Ali got angry with
Ay~e.
b. Ahmet Ali'yi Ayse'ye kizdud1/ otkelendirdi/ sinirlendirdi.
Causer Experiencer Theme
(NOM) (ACC)
(OAT)
Ahmet made Ali angry at Ay~e .
c. Ali Ayse'yi kizdud1/ otkelendirdi/ sinirlendirdi.
Causer Experiencer
(NOM) (ACC)
Ali made Ay~e angry.
There are two ways of looking at the verb forms in (6c), (7c) and (8c). From
the first point of view (c) sentences imply a deleted dative reflexive pronoun coindexed
with the Causer:
(8) Ali Ayse'yi (kendine) danltt1/ gucendirdi/ klistlirdU/ kizdirdi/ otkelendirdi/
Causer Experiencer
(NOM) (ACC)
Theme
(DA T)
sinirlendirdi/ ozendirdi.
From the second point of view, these verbs are lexical causatives of Type 2
where the subject is interpreted either as a Causer or as a Theme:
(9) Ali Ayse'yi dariltt1/ gucendirdi/ kustlirdU/ kizd1rdii ofkelendirdi/
Causer/ Theme Experiencer
(NOM)
(ACC)
sinirlendirdi/ ozendirdi.
The following rules can be suggested for the causativization of Type 3a psych
verbs. The first two rules are invalid for some Type 3a verbs as mentioned above:
Type 3a Psych Construction
Experiencer + Theme
(NOM)
(OAT)
Type 3a ~ausative Psych Construction
Causer+ Experiencer+ Theme
(NOM)
(ACC)
(OAT)
Causer+ Theme+ Experiencer
(NOM)
(OAT)
(ACC)
129
*Causer+ Theme
(NOM)
(OAT)
Causer+ Experiencer
Table 9.
(NOM)
(ACC)
Type Type 3a Causat1ve Psych Construction
II. 3. 2. 1. 4. Type 3b Psych Verbs and Morphological Causativity
The most thoroughly discussed phenomenon in the literature of psych verbs is
the psychological causativity and the representative verbs fear/ frighten which are kork-1
korkut- in Turkish. Goksel (1993) claims that these verbs are not different from ordinary
one place predicates. The only difference is in their causative form which results from the
ablative marking in Turkish. She claims that the ungrammaticality of such a sentence like:
(1) *Zehra beni Nam1k'tan korkuttu.
is because of the impossibility of using a psych causative with an ablative which itself
expresses a kind of causality. According to her, not only for psych causatives but also for
all causatives, it is ungrammatical to use causal ablative arguments in causative
constructions. Therefore she regards (2) as ungrammatical as well as (1 ):
(2) *Tetanoz John'u hastahktan oldiirdii.
*Tetanus killed John with disease.
In our analysis of Type 3b psych verbs, such examples are not ungrammatical
since it is not uninterpretable to regard Zehra/ Tetanoz as the Causer, benil John 'u as the
Experiencer and the ablative marked Namtk 'tan/ hastalzktan as the Theme. It is possible to
construct a causal chain in which one Causer causes the other Causer which causes another
event as in the following example:
(3) Kirli su Ay:;;e'yi dizanteriden oldiirdii.
Causer Experiencer
Theme
(NOM)
(ACC)
(ABL)
Dirty water killed Ay~e with dysentery.
130
It is also possible for an animate outside Causer to cause an Experiencer
experience a psychological state. In this case, there are both a Causer argument and an
ablative Theme argument which is a more direct cause of the emotional state of the
Experiencer:
(4) Annesi cocugu doktordan korkuttu.
Causer
Experiencer
Theme
(NOM)
(ACC)
(ABL)
Her mother made the child fear the doctor.
The preverbal focus makes the sentence sound more acceptable:
(5) Cocugu doktordan annesi korkuttu.
Experiencer
Theme
Causer
(ACC)
(ABL)
(NOM)
In all such examples in this study, the Theme expresses the direct Causer of the
event and the Causer expresses a secondary, indirect or an outside Causer.
The majority of the verbs of Type 3b can be causativized:
(+) CAUS.
bez -(Dir)h1k -(Dir)bunal-(t)c;ek -(Dir)igren -(Dir)
kork -(t)ku~kulan -(Dir)sogu -(t)~iiphelen -(Dir)tiksin -(Dir)usan -(Dir)utan -(Dir)-
ONLYCAUS.
0
DIFF.PSYCH
0
NON-CAUS.
0
(-)CAUS.
slkll -(Dir)-
..
Table 10. Type 3b Psych Verbs and CausatlVIty
As the following examples (1)- (3) shows, a nominative or accusative Causer
is added, Experiencer is accusative marked and the ablative marked Theme stays the same.
However, (b) sentences become much more acceptable when the Causer follows the
Theme as the Experiencer in the subject position ((c) sentences) due to the focusing of the
preverbal element. (d) and (e) alternatives are not ungrammatical but in (d)s, there is not a
specific source of the mental/ emotional state of the Experiencer, therefore, there is a
131
generality. In (e)s, there is not a specific Experiencer of the mental! emotional state and,
therefore, a generic Experiencer is interpreted, so, the optional deletion does not cause
ungrammaticality but only an information gap.
Ay~e'den
(1) a. Ali
Experiencer
Theme
(NOM)
(ABL)
btktt/ usand1.
Ali was fed up with
Ay~e.
b. Bu davran1~lar Ali'yi
Ay~eden
Causer
Experiencer
Theme
(NOM)
(ACC)
(ABL)
btktudt/ usand1rd1.
Such behaviors ofherselfmade Ali fed up with
Ay~e.
c. Ali'yi Ay~e'den bu davran1~lar btkttrdt/ usand1rd1.
Experiencer
Theme
(ACC)
(ABL)
Causer
(NOM)
d. Bu davrant~lar Ali'yi btktudt.
Causer
Experiencer
(NOM)
(ACC)
(generally)
e. Bu davrant~lar Ay~e'den btktlrd1. (everybody)
Causer
Theme
(NOM)
(ABL)
(2) a. Ali tavuktan igrendi/ tiksindi.
Experiencer
Theme
(NOM)
(ABL)
Ali was disgusted with chicken.
b. Bu lokanta Ali'yi tavuktan igrendirdi/ tiksindirdi.
Causer
Experiencer Theme
(NOM)
(ACC)
(ABL)
This restaurant made Ali disgusted with chicken.
Experiencer
Theme
Causer
(ACC)
(ABL)
(NOM)
d. Bu lokanta Ali'yi igrendirdi/ tiksindirdi. (generally)
Causer
Experiencer
(NOM)
(ACC)
e. Bu lokanta bahktan igrendirdi/ tiksindirdi.
Causer
Theme
(NOM)
(ABL)
(3) a. Ali
Experiencer
(NOM)
Ay~e'den ku~kulandt/ ~uphelendi.
Theme
(ABL)
Ali was suspected of Ay~e.
(everybody)
132
b. Polis Ali'yi Ay~e'den ku~kulandirdi/ ~iiphelendirdi.
Causer Experiencer Theme
(NOM)
(ACC)
(ABL)
The police made Ali suspected of Ane.
c. Ali'yi Ayse'den polis ku~kulandudii ~iiphelendirdi.
Experiencer
Theme
Causer
(ACC)
(ABL)
(NOM)
d. Polis Ali'yi ku~kulandud1. (generally)
Causer Experiencer
(NOM)
(ACC)
e. Polis Ay~e'den ku~kulandud1.
Causer
Theme
(NOM)
(ABL)
(everybody)
The (d) examples from (1) to (3) demonstrate a similar case with the cases in
other verb types. The causative forms of these verbs can either be regarded as having a
deleted optional ablative Theme or they can be regarded as separate entries which belong
to Type 2:
( 1) Ali Ayse 'yi bikt1rdi/ igrendirdi/ tiksindirdi/ ku~kulandud1/ ~iiphelendirdi/
korkuttu/ iirkiittii/ utand1rd1.
The latter treatment of these verbs (as Type 2 verbs) correspond to the
universal treatment since in all of the languages analyzed in the psych verb literature
(English, Italian, French, Spanish, Finnish, Polish, Greek etc.), these verbs belong to the
second class named as frighten type verbs.
In the examples above the ablative expresses the source of the mentaV
emotional state. Therefore, the rules for Type 3b causativization can be stated as:
133
II. 3. 2. 1. 5. Type 4 Psych Verbs and Morphological Causativity
(+)CAUS.
az -(Dir)bilin9len -(Dir)co~ -(Dir)eglen -(Dir)heyecanlan -(Dir)kudur -(t)sakinle~ -(Dir)yatl~ -{Dir)-
0
ka9 -(Ir)-
dol-(Dir)don -(Dir)kabar -(t)-
ap1~ -(Dir)durul-(t)kubar -(t)siiziil-(t)yirtll-(t)-
Table 12. Type 4 Psych Verbs and Causativity
The only argument of the Type 4 verbs is the Experiencer. Causativization adds
one more argument to the following intransitive sentences which is a nominative Causer.
Therefore causativization process turns all of the intransitive psych verbs of Type 4 into
transitive Type 2 verbs:
( 1) a. Cocuklar azd1.
Experiencer
(NOM)
The children got wild.
b. Cocuklar az1tt1.
Experiencer
(NOM)
The children got wild.
c. Ablast cocuklart azdud1.
Causer
(NOM)
Experiencer
(ACC)
Their sister made the children out of control.
(2) a. Kalabahk co~uyor.
Experiencer
(NOM)
The crowd cheers.
b. Sarktct kalabahg1
Causer
(NOM)
co~turuyor.
Experiencer
(ACC)
The singer cheers the crowd.
(3) a. Ay:;;e heyecanlamyor.
Experiencer
(NOM)
Ay~e gets excited.
134
b. Ali Ayse'yi heyecanlandmyor.
Causer Experiencer
(NOM) (ACC)
Ali makes Ane excited.
(4) a. Annem sakinle~iyor/ yat1~1yor.
Experiencer
(NOM)
My mother calms down.
b. Baham annemi
Causer
(NOM)
sakinle~tiriyor/ yat1~tlnyor.
Experiencer
(ACC)
My father makes my mother calm down.
The following formula summarizes the only way of causativizing the Type 4
verbs:
Type 4 Psych Construction
Experiencer
(NOM)
Type 4 Causative Psych Construction
Causer+ Experiencer
(ACC)
(NOM)
Table 13. Type 4 Causative Psych Construction
In this section, we analyzed the causativity properties of Type 1, 2, 3a, 3b and 4
psych verbs in Turkish. The analysis has shown that in a causative construction, some
verbs which belong to other verb classes behave like lexical causatives of Type 2. In all of
these verbs the subject can either be regarded as a Causer or as a Theme as the Table 14
shows. On the other hand, the following Table 1 on pages XIX and XX summarize the
causativity properties of all psych verb types in Turkish:
~
Experiencer + Theme+ Causer
(NOM)
(ABL) (ACC)
TYP~4
Experiencer
(NOM)
Ali'yi
tavuktan bu lokanta igrendirdi.
Experiencer Theme
Causer
(ACC)
(ABL)
(NOM)
(This restaurant made Ali disgusted with
chicken.)
'
Causer/ Theme+ Experiencer
Ahmet
Ali'yi utandtrdt.
(NOM)
(ACC)
Causer/ Theme Experiencer
(NOM)
(ACC)
=Type2
(Ahmet embarrassed Ali.)
Causer+ Theme
Bu lokanta bahktan igrendirdi.
(NOM) (ABL)
Causer
Theme
(NOM)
(ABL)
(This restaurant made Ali disgusted .)
Causer+ Experiencer
Baham annemi sakinle~tiriyor/ yatt~tmyor.
(NOM)
(ACC)
Causer Experiencer
(NOM) (ACC)
==Type 2
(My father makes my mother calm down .)
Table I. Summary of the Causativity Properties of Turkish Psych Verbs
'
135
Experiencer
(NOM)
(ACC)
r.:-------1 lctistiirdti/ ktzdudt/ ofkelendirdi/ ozendirdi/ sinirlendirdi/ ~a~trttl.
Causer/ Theme
Experiencer
(NOM)
(ACC)
~---,-------1 ku~kulandudt/
~Uphelendirdi/
tiksindirdi/Urktittti/
r.---:--.--i utandudt.
Type4
Babam
Causer/ Theme
(NOM)
annem1
Experiencer
(ACC)
yatl~tmyor.
usandtrdt/
136
Periphrastic Causative Psych Verbs
II. 3.2. 2.
In Turkish, causativity can be expressed with periphrastic constructions as well
as by lexical means. Morphological causativity is a verbal process though periphrastic
causativity is a nominal one. Sag/a- and neden ol- are the most often used predications of
these constructions.
With non psych verbs the DAT marked Actor/ Doer/ Agent of the non
causative construction gets the GEN marking and the causation event with the nominalized
verb (with -rnA) takes the accusative case for sagla- (1b) and dative case for neden o (1c):
( 1) a. Ali Ay$e'ye kitab1 okuttu.
Causer
Causee
Theme
(NOM)
(DA T)
(ACC)
b. Ali Ay$e'nin kitab1 okumasm1 saglad1.
Causer
Causee
Theme
(NOM)
(GEN)
(ACC)
c. Ali Ay§e'nin kitab1 okumasma neden/ neden oldu.
Causer
Causee
Theme
(NOM)
(GEN)
(DA T)
Stylistic variations changing the syntactic position of the arguments are also
possible.
This section briefly analyzes the interrelationships between the psych verb
types, morphological causativity and periphrastic causativity in Turkish.
II. 3. 2. 2. 1.
Type 1 Psych Verbs and Periphrastic Causativity
As mentioned, Causer+ Experiencer order was ungrammatical for Type 1
(NOM.)
(DA T.)
137
psych verbs without the existence of an accusative Theme . Since a periphrastic causative
construction does not alter the thematic roles, the ungrammaticality of (la) can not be ruled
out by (lb):
(1) a. *Ali Ahmet'e affettirdi.
b. *Ali Ahmet'e affetmesine neden oldu.
All of the verbs of Type 1 which can be morphologically causativized can also
be used in periphrastic constructions. As mentioned in the previous section, except for two
verbs, all of the causativizable Type 1 verbs have similar semantic interpretations in the
periphrastic causative construction:
Ay~e'yi
(2) a. Ali
Causer
(NOM)
Theme
(ACC)
Ali made
Ay~e
b. Ali
unutturdu.
be forgotten.
Ay~e'yi
herkese unutturdu.
Experiencer
(OAT)
Ay~e'nin
c. *Ali
d. Ali
Ay~e'nin
unutmasm1 saglad1/ neden oldu.
(herkes tarafmdan) unut(ul)masm1 sagladt/ neden oldu.
Unlike the other verbs Type 1, the two exceptional verbs of this type which
behave like Type 2 verbs do not have passive interpretations with an optional Experiencer
by phrase in their causative forms:
(3) a. Ali
Ay~e'yi dti~Undtirdli.
Causer Experiencer·
(NOM) (ACC)
Ali made Ay~e think.
b. *Ali
Ay~e'yi
Causer Experiencer
(NOM) (ACC)
c. Ali
herkese
dti~Undtirdli.
Theme
(DA T)
Ay~e'nin dti~Unmesini
saglad1/ neden oldu.
d. *Ali Ay~e'nin (herkes tarafmdan) dli~Un(Ul)mesini saglad1/ neden oldu.
138
(4)a. Ali Ane'yi ktskand1rd1.
Causer Experiencer
(NOM) (ACC)
Ali made Ane envy.
b. *Ali
Ay~e'yi
herkese k1skandtrd1.
Causer Experiencer
(NOM) (ACC)
Theme
(DA T)
c. Ali Ay~e'nin ktskanmasmt sagladt/ neden oldu.
b. *Ali Ay~e'nin (herkes tarafmdan) klskan(ll)masm1 sagladt/ neden oldu.
Type 1 verbs which can not be morphologically causativized are acceptable in
periphrastic causative constructions (although they are not perfect). The observed tendency
is to use mental verbs often with sag/a- as in (Sd):
(S)a. Ali bu sorunu ongordli.
Experiencer
(NOM)
Theme
(ACC)
Ali anticipated this problem.
b.* Ahmet Ali'ye bu sorunu ongordlirdli.
Causer Experiencer
(NOM)
(DA T)
Theme
(ACC)
Ahmet made Ali anticipate the problem.
c. ?Ahmet Ali 'nin bu sorunu ongormesine neden oldu.
Causer Experiencer
(NOM)
(GEN)
Theme
(ACC)
d. Ahmet Ali 'nin bu sorunu ongormesini saglad1.
Causer Experiencer
(NOM)
(GEN)
Theme
(ACC)
On the other hand, emotional verbs tend to be used with neden ol rather than
sag/a:
(6) a.
Ay~e
Ali'yi arzulad1.
Experiencer Theme
(NOM) (ACC)
Ay~e
desired Ali.
b. Ahmet Ay~e'ye Ali'yi arzulatt1.
Causer Experiencer
(NOM)
(DA T)
Ahmet made
Ay~e
Theme
(ACC)
desire Ali.
c.? Ahmet Ay~e'nin Ali'yi arzulamasm1 saglad1.
Causer
(NOM)
Experiencer
(GEN)
Theme
(ACC)
139
c. Ahmet Ay$e'nin Ali'yi arzulamasma neden oldu.
Causer
(NOM)
Experiencer
(GEN)
Theme
(ACC)
The underlying semantic reason may be due to the semantic difference between
sag/a- and neden ol-. The former implies a conscious effort of the Causer to make the
Experiencer experience the psych state. On the other hand the latter implies a nonconscious
act (or not an act at all) of the Causer.
This argument may be linked to the assumption that emotional activities are
less controllable than the mental ones. Therefore, mental leanings may be controllable by
outside agents by physically or consciously doing something but it is not easy for an
outside Causer to have the control over the emotional inclinations of someone. Only the
existence of a Causer or something that it does unconsciously causes the Experiencer to
experience an emotional state.
II. 3. 2. 2. 2.
Type 2 Psych Verbs and Periphrastic Causativity
As we mentioned in the previous sections, all of the Type 2 verbs are
inherently causative and they can not be further causativized. That means all of the Type 2
verbs are lexical causatives and they can not be causativized morphologically:
(1) a. Konu$mact dinleyicileri baydt/ s1kt1.
Theme
(NOM)
Experiencer
(ACC)
The speaker bored the audience.
b. *Adam konu$mactya dinleyicileri baydtrdt/ stktlrdt.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*The man made the speaker bore the audience.
However, it is possible to use all of the Type 2 verbs in periphrastic causative
constructions with the verb passivized and with a reflexive interpretation as in (2d) and
140
(3d). That demonstrates our claim that the restriction on the causativity of these verbs is
neither phonological nor semantic but a morphological one:
(2) a. Ane Ali'yi UzdU.
Theme Experiencer
(NOM) (ACC)
Ay~e depressed Ali.
b. *Ahmet Ali'ye Ayse'yi UzdUrdU.
Causer
(NOM)
Theme
(DAT)
Experiencer
(ACC)
*Ahmet made Ali depress Ay~e.
c. Ahmet Ali'nin Ay~e'yi Uzmesine neden oldu.
d. Ayse Ali 'nin Uzlllmesine ned en oldu.
Causer Experiencer
(NOM) (GEN)
(3)
a.~
Ali'yi ktrd1.
Theme Experiencer
(NOM) (ACC)
Ay~e
hurt Ali.
b. Ahmet Ali'ye Ayse'yi ktrdtrdL
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
*Ahmet made Ali hurt Ay~e.
c. Ahmet Ali'nin Ay~e'yi ktrmasma neden oldu.
c. Ali Ay:;;e'nin kmlmasma neden oldu.
Causer Experiencer
(NOM) (GEN)
As mentioned, some of the Type 2 verbs seem to be inseparable from their
causative morphemes. If we accept the
-t
morpheme attached to these verbs as the
causative morpheme, the semantic interpretations of the following periphrastic causatives
are reflexive:
(3) a. *Ali Ay:;;e'yi kt~ktrdt.
Theme Experiencer
(NOM) (ACC)
Ali incited Ay~e.
b. *Ahmet Ali'ye Ay:;;e'yi kt~ktrttl.
Causer
(NOM)
Theme
(DA T)
Ahmet made Ali
Experiencer
(ACC)
incite Ay~e.
141
c. Ali Ay~e'yi kt~ktrtti ..
Theme Experiencer
(NOM) (ACC)
Ali incited Ay~e.
d. Ali Ay~e'nin kt~kmnasma neden oldu.
Causer Experiencer
(NOM)
(GEN)
(4) a. *Ali Ay~e'yi aydmlad1.
Theme Experiencer
(NOM) (ACC)
b.* Ahmet Ali'ye Ay~e'yi aydmlatt1.
Causer
(NOM)
Theme
(DA T)
Experiencer
(ACC)
Ahmet made Ali enlighten
c. Ali
Ay~e'yi
Ay~e.
aydmlatt1.
Theme Experiencer
(NOM) (ACC)
Ali enlightened Ay~e.
d. Ali
Ay~e'nin
aydmlanmasm1 saglad1.
Causer Experiencer
(NOM)
(GEN)
II. 3. 2. 2. 3.
Type 3a Psych Verbs and Periphrastic Causativity
All of the Type 3a verbs that can be causativized can also be used m
periphrastic psych constructions:
(1) a. Ali Ay~e'ye danldtl glicendi/ ktistti.
Experiencer Theme
(NOM)
(DAT)
Ali put out with Ane.
b. Ahmet Ali'yi Ay§e'ye danlttt/ glicendirdi/ ktistlirdli.
Causer Experiencer Theme
(NOM) (ACC)
(DAT)
Ahmet made Ali.
c. Ali Ay§e'yi danlttt/ glicendirdi/ klisttirdli..
Causer Experiencer
(NOM) (ACC)
d. Ali Ay§e'nin danlmasmal glicenmesine/ klismesine neden oldu.
Causer Expeeriencer
(NOM)
(GEN)
142
The only Type 3a verb aldzr which is not a psych verb without the causative
morpheme can not be used in a periphrastic psych construction. However all of the Type
3a verbs whose non causative forms are psych are acceptable as periphrastic causatives:
(1) a. AyO?e bu elbiseye baylld1. (psych)
Experiencer
(NOM)
Ay~e
Theme
(DAT)
liked this dress greatly.
b. *Ali AyO?e'yi bu elbiseye baylltt1.
Causer Ezxperiencer
(NOM) (ACC)
Ay~e
Theme
(DA T)
made Ali like this dress.
c. Ali AyO?e'nin bu elbise'ye bayllmasma neden oldu.
Causer Experiencer
(NOM) (GEN)
Theme
(DAT)
There are a few number of verbs of Type 3a which can not be causativized.
These verbs can perfectly be causativized in a periphrastic construction:
(3) a. *Ahmet Ali'yi AyO?e'ye giivendirdi.
Causer Experiencer
(NOM) (ACC)
Theme
(DA T)
Ahmet made Ali trust Ay~e.
b. Ahmet Ali'nin AyO?e'ye giivenmesini saglad1.
Causer Experiencer
(NOM)
(GEN)
Theme
(DA T)
(4) a. *Ahmet Ali'yi AyO?e'ye kmldtrdl.
Causer Experiencer
(NOM) (ACC)
Theme
(DA T)
b. Ahme! Ali'nin AyO?e'ye kmlmasma neden oldu.
Causer Experiencer
(NOM)
(GEN)
II. 3. 2. 2. 4.
Theme
(DAT)
Type 3b Psych Verbs and Periphrastic Causativity
All Type 3b verbs can be used in periphrastic psych constructions. Most of the
verbs of Type 3b that we examine have some kind of negative semantic content on the part
of the Experiencer and they sound much more acceptable with neden ol- which led us to
143
notice that neden ol- is more compatible with negative psych events while sag/a-
IS
compatible with positive ones:
(1) a. Bu davram~lar Ali'yi Ay~eden btktudt/ usandtrdt.
Causer
(NOM)
Experiencer
(ACC)
Theme
(ABL)
Such behaviors ofherselfmade Ali fed up with
Ay~e.
b. Bu davrant~lar Ali'nin Ay~e'den btkmasma neden oldu.
Causer
(NOM)
Experiencer
(GEN)
Theme
(ABL)
There is only one Type 3b verb which can not be causativized; szkzl. This verb
can also be causativized in a periphrastic construction:
(2) a. *Ahmet Ali'yi
Ay~e'den
Causer Experiencer
(NOM)
(ACC)
sikildirdt.
Theme
(ABL)
b. Ahmet Ali'nin Ay~e'den stkilmasma neden oldu.
Causer Experiencer Theme
(NOM)
(ACC)
(ABL)
II. 3. 2. 2. 5.
Type 4 Psych Verbs and Periphrastic Causativity
Type 4 verbs that can be morphologically causativized can also be causativized
periphrastically. Other groups of Type 4 verbs whose either non causative or only the
causative forms are psych can also be used periphrastically. The non causativizable verbs
of Type 4 can be causativized too if used in periphrastic constructions (although they do
not sound perfectly acceptable):
(1) a. Ali duruldu.
Experiencer
(NOM)
Ali calmed down.
b. *Ay~e Ali 'yi durulttu.
Causer Experiencer
(NOM) (ACC)
Ane made Ali calm down.
144
c. Ay$e Ali 'nin durulmasma neden oldu.
Causer
Experiencer
(NOM)
(GEN)
This brief analysis shows that first, all types of psych verbs in Turkish can be
used in periphrastic constructions no matter whether they are lexically or morphologically
causative.
Second, it became clear that all of the verbs of Type 1, 2, 3 and 4 which can not
be causativized for this or that reason, can be causativized in periphrastic constructions.
Third, it seems that sag/a- constructions mostly refer to positive psych events
on the part of the Experiencer and neden ol- constructions refer to rather negative ones.
Lastly, it is anticipated that sag/a- constructions with the implication of a
conscious effort are more compatible with mental events while neden ol- constructions
which imply a non conscious/ accidental cause are more compatible with emotional ones.
II. 3. 2. 3.
Causativity and the Permissive Reading
As exemplified in the section above, all types of psych verbs that can be
morphologically causativizable and non-causativizable can be used in periphrastic psych
constructions. The periphrastic causative is a kind of paraphrase of the morphological
counterpart just as it is the case with non-psych causatives. When the morphological
counterpart is negated, the periphrastic paraphrase has a permissive reading:
(1) a. Ailesi
Ay~e'yi
okuttu.
b. Ailesi Ay~e'yi okutmad1.
c. Ailesi
Ay~e'nin
okumasma izin vermedi.
Her parents did not let Ay~e go on her education.
145
With a few number of psych verbs, this permissive reading is more significant
than the others. Ignoring the thematic relations, in the following sentences, (c)s are the
periphrastic permissive paraphrases of the morphological (b)s. For the purposes of focus
and emphasis, the dative argument can move to the preverbal position:
(2) a. Ali bize Ay$e'yi unutturdu.
Causer Experiencer Theme
(NOM) (DA T) (ACC)
Ali made us forgive
Ay~e.
b. Ali bize Ay§e'yi unutturmadt.
Causer Experiencer Theme
(NOM) (DA T) (ACC)
c. Ali Ay§e'nin (bizim taraf1m1zdan) unutulmasma izin vermedi.
Causer
(NOM)
Theme
(GEN)
Ali did no let Ane to be forgotten (by us).
(3) a. Ali Ay§e'yi bize kUstiirdU.
Causer Experiencer Theme
(NOM) (ACC)
(DA T)
Ali made Ay~e put up with us.
b. Ali Ay§e'yi bize kUstUrmedi.
Causer Experiencer Theme
(NOM) (ACC) (DAT)
c. Ali Ayse'nin bize ktismesine izin vermedi.
Causer Experiencer Theme
(NOM)
(GEN)
(DA T)
Ali did not let Ay~e put up with us.
In the following example, a Type 2 verb which is assumed to be an inherently
causative verb can not be causativized (4a) but the negative form (4b) seems to be
acceptable with a corresponding permissive periphrastic paraphrase (4c):
( 4) a. *Ali Ayse'yi herkese iizdUrdU.
Causer Experiencer
(NOM) (ACC)
Ali made
Ay~e
Theme
(OAT)
be depressed by everyone.
b. Ali Ay§e'yi kimseye Uzdtirmedi.
Causer Experiencer
(NOM) (ACC)
Theme
(DA T)
c. Ali Ay§e'nin (kimse tarafmdan) Uztilmesine izin vermedi.
Causer Experiencer
(NOM)
(GEN)
Ali did not let anyone depress
Ay~e.
146
II. 3. 2. 4.
Control, Causativity and the Psych Events
As mentioned and tested before, emotional activities seem to be less
controllable than the mental ones. Emotional activities are generally regarded vaguely as
personal matters which can not be induced or created in other people. Control can be
considered as two types with respect to the controller:
1. intemaV self control
2. external/ outer control
In fact, the semantic distinction between the two seems not to create a structural
distinction. That means if a psych event is not self controllable, it can not be controlled by
an outer agent too and vice versa. Therefore, the point is being either controllable or noncontrollable.
Imperative test can be used as a diagnostic to determine the controllability of
psych events. The analysis shows that all of the psych verbs in their imperative form
require either an adverbial or an NP marked with the case that the verb normally requires.
However, the test is not sensitive to the type of the psych verb (and therefore not to the
case of the complement) but rather sensitive to the meaning of the verb. Since imperative
suggests a kind of order/ advice and since an advice generally implies a positive or useful
suggestion on the part of the addressee, psych verbs with a positive meaning can be used in
an imperative form:
(1) iyice ogren.
Learn well.
(2) Yaptiklarmdan utan.
Be ashamed of what you did.
(3)
Ba~artlannla
gururlan.
Be proud of your successes.
147
(4) Amlarla avun.
Console yourself with the memories.
(5) Biraz huslan/ ne~elen.
Cheer a little.
Most of the other psych verbs with a positive meaning can also be used in the
imperative form but with a different meaning (which is more controllable than the original
psych meaning):
(1) Tannya inan.
Believe in God.
Accept the existence of the God.
(2) Ulkeni sev.
Love your country.
Be devoted to your country.
(3) Olumlu
dti~tin.
Think positive.
Look at it in a more positive way.
(4) Once ba~armay1 iste.
First, want to succeed.
First, intend to succeed.
(5) Bize ac1.
Feel pity for us.
Be good to us.
(6) istenmedigini anla.
Understand that you are notwanted.
Realize that you are not wanted.
(7) Sevildigini bil .
Know that you are loved.
Realize that you are loved.
(8) Bu kopekten kork.
Be afraid of this dog.
Watch out this dog.
When the verb has a negative meaning, the imperative should be negated too
(which will mean a positive advice):
(9) a. ?Kocam k1skan.
Envy your husband.
148
b. Kocam ktskanma.
Do not envy your husband.
(10) a. ?<;ah~maktan b1k.
Get sick of studying.
b.
<;ah~maktan
b1kma.
Do not get sick of studying.
( 11) a. ?Bahktan igren.
Disgust with the fish.
b. Bahktan igrenme.
Do not disgust with the fish.
It may be possible to use some non controllable emotional events in the
imperative form with the implication that a physical act or a series of physical acts are
done to evoke the emotional state on the Experiencer:
(9) Onu Oz.
Make him depressed (by doing something/ saying something etc.)
In fact iiz- is inherently a causative verb and the controllability phenomenon
with this meaning is closely related to causativity. If the psych event is a controllable
event, it can be causativized but if the event is a non-controllable one, the causative form
seems rather odd. If the event is a non controllable one but there is still a causative form,
this implies that there is a physical act or a series of physical acts involving in the causative
event to evoke the emotional state on the Experiencer:
At this point there is another distinction; an Experiencer can experience the
·emotional state because of either of the following:
1. a conscious or an unconscious activity of someone or something or
2. only the existence of someone or something
causes the psych state in the Experiencer
149
In English most of the psych verbs which are stative can not be used
progressively and therefore, (10) may have two possible interpretations; (a) referring to a
habitual disposition and (b) referring to the moment of speech:
(1 0) !_fear the dog.
Experiencer
Theme
a. The existence of a dog frightens me even if it does do anything to me at
all/ I always fear the dogs/ I have a phobia of dogs.
b. The dog is looking at me and barking and this frightens me (at that
moment).
However in Turkish, the habitual psych state is used with the aorist and there is
not a restriction on the progressiveness of stative psych events:
(11) a. Kopekten korkar1m.
(habitual psych state)
The existence of a dog frightens me even if it does do anything to me at all/
I always fear the dogs/ I have a phobia of dogs.
b. Kopekten korkuyorum.
(moment of speech)
The dog is looking at me and barking and this frightens me (at that
moment).
In the causative constructions, when there is not an aorist or the progressive to
indicate the habituality or the moment of speech, (12) is unacceptable with the
interpretation (a) but acceptable with the interpretation (b):
(12) Annem beni doktordan korkuttu.
Causer Experiencer Theme
(NOM)
(ACC)
(ABL)
a. *My mother made me feel the emotional state of fearing the doctor.
b. My mother said something/ did something (consciol,lsly or unconsciously)
to make me fear the doctor.
In (10) and (11), the dog is the direct cause of the fear. However, in (12) the
mother is the indirect Causer but the doctor is the direct Causer of the fear. It is enough for
a direct Causer just to exist to make the Experiencer experience the emotional state. On the
150
other hand, the existence of an indirect Causer is not enough as a condition to make the
Experiencer experience the emotional state. He/ she has to do something physically for the
psych event to occur as the unacceptability of (12a) indicates.
The ambiguity can be resolved if annem is emphasized as a Causer occupying
the preverbal position:
(13) a. Annem beni doktordan korkuttu. (Ambiguous)
Causer Experiencer Theme
(NOM)
(ACC)
(ABL)
b. Beni doktordan annem korkuttu. (Non-ambiguous)
Experiencer
(NOM)
II. 3. 3.
Theme
(OAT)
Causer
(NOM)
Psych Verbs and Passivization in Turkish
Passivization process in Turkish deletes or absorbs the subject argument which
can either be expressed by an optional tarafindan (by phrase) or left unexpressed.
Therefore, passive is referred to as a thematic role absorber or intransitivizer. In the case
of psych verbs, it is the Experiencer argument that is optionally delete.d and it is only the
Theme argument that survives in the surface:
DELETED
ARGUMENT
SURVIVIG
.
ARGuMENT
TRANSITIVE PSYCH VERBS.
'
Subject or Object Experiencer Argument
" Subject or Object Theme Argument
Table I.
..
Passtvtzatton m Psych Verbs
(1) a. Ali Ayse'yi begendi.
Experiencer Theme
(NOM)
(ACC)
Ali liked Ane.
b. Ayse (Ali tarafmdan) begenildi.
Theme
(NOM)
Ay~e was
Experiencer
liked by Ali.
OTHER TRANSITIVE VERBS
Subject Argument
Object Argument
151
As a result of the multi functionality of the inflectional and derivational suffixes
in Turkish, there are cases in which some forms do not correspond to specified semantic
functions. This is significant especially in the case of reflexive and passive morphemes. As
argued by Tietze (1989) one form may have the semantic function of the other. Thus, this
two phenomena need to be analyzed with reference to each other.
II. 3. 3. 1.
Type 1 Psych Verbs and Passivization
The phonological conditions determine the attachment of either-(!)/ or -(l)n to
give a passive meaning.-(!)/ is attached to the ones which end with a consonant:
VERB
begendU~Un-
ktskansansevurnunutvarsay-
MORPHEME
FUNCTION
-II
-II
passive
passive
passive
passive
passive
passive
passive
-Il
-II
-II
-II
-II
-II
Table 2.
passive
Type I Psych Verbs With -II
and -(l)n is attached to others which end either with a vowel or a lateral:
VERB
arzulaalgtlabellebilhat1rlakavraozleumursa-
MORPHEME
FUNCTION
-(I)n
-(I)n
-(I)n
-(I)n
passive
passive
passive
passive
passive
passive
passive
-(l)n
-(I)n
-(I)n
-(I)n
Table 3.
passive
Type I Psych Verbs With (l)n
152
Among these, only the mental verb an/a- is passivized not as anla(n)- but as
anla($)(Z/}-:
(1) a. Ali Ane'nin niyetini anladt.
Experiencer
Theme
(NOM)
Ali
(ACC)
understood Ay~e's
intention.
b. *Ay~e'nin niyeti (Ali tarafmdan) anlandt.
c. Ayse'nin niyeti (Ali tarafmdan) anla~aldt.
Theme
Experiencer
(NOM)
Most of the time it is the case that if the verb has primarily a non psych
meaning, all of the voice suffixes give their original meaning to the first sense of the verb.
Therefore, the passive counterparts of the following verbs are non psych:
VERB
9ak- "to understand"
91kar- "to understand"
ktvtr- "to understand"
kur- "to imagine"
salla- "to care"
Table 4.
MORPHEME. .. .••.•• ..
-11
-11
-11
-11
-(l)n
fi
PASSIVE
9aktl- "to be nailed"
91kanl- "to be taken out"
ktvnl- "to curl"
kuru!- "to settle"
sallan- "to swing"
Type I PasslVe non psych Verbs
Sometimes both the active and the passive counterparts have psych meanings
but they are different from each other:
VERB
96z- "to solve"
kap- "to understand"
tak- "to care"
Table 5.
II. 3. 3. 2.
PASSIVE
96ziil- "to relax"
kaptl- "to give oneself to smth."
taktl- "to kid"
Type I Passtve and Active Psych Verbs
Type 2 Psych Verbs and Passivization
Except for two verbs biiyiile- and etkile-, all of the other verbs of Type 2 take
the passive morpheme -(I)/ and have a reflexive meaning. However, these reflexive
counterparts can also be regarded as passive constructions with deleted by phrases:
153
(1) a.
Ay~e
beni biiyiiledi.
Theme Experiencer
(NOM) (ACC)
Ay~e charmed me.
b. ?Ben
(Ay§e tarafmdan) biiyiilendim.
Experiencer
Theme
(NOM)
Ay~e.
I was charmed by
c. *Ben Ayse'ye/ Ayse'den btiyiilendim.
Experiencer
Theme
(NOM)
(DA T/ ABL)
(2) a. Ayse beni etkiledi.
Theme Experiencer
(NOM)
Ay~e
(ACC)
attracted me.
b. ?Ben (Ay~e tarafmdan) etkilendim.
Experiencer
Theme
(NOM)
I was attracted by
Ay~e.
c. Ben Ay§e'den etkilendim.
Experiencer Theme
(NOM)
(ABL)
I was attracted by
Ay~e .
(3) a. Ane beni baydt.
Theme Experiencer
(NOM) (ACC)
Ay~e
bored me.
b. ?Ben (Ay§e tarafmdan) baytldtm.
Experiencer
Theme
(NOM)
?I was bored by Ay~e.
c. Ben Ayse'ye baytldtm.
Experiencer Theme
(NOM)
(OAT)
I loved Ay~e very much.
(4) a. Ayse beni bogdu.
Theme Experiencer
(NOM)
Ay~e
(ACC)
annoyed me.
b. ?Ben (Ay~e tarafmdan) boguldum.
Experiencer
Theme
(NOM)
I was annoyed by
Ay~e.
154
c. *Ben Ayse'ye/ Ayse'den boguldum.
Experiencer
Theme
(NOM)
(OAT/ ABL)
(5) a. Ayse beni bozdu.
Theme Experiencer
(NOM) (ACC)
Ay~e embarrassed
me.
b. ?Ben (Ayse tarafmdan)bozuldum.
Experiencer
Theme
(NOM)
I was embarrassed by Ay~e.
c. Ben Ayse'ye bozuldum.
Experiencer
Theme
(NOM)
(OAT)
I felt disconcerted because of Ane.
(6) a. Ayse beni ktrdt.
Theme Experiencer
(NOM)
Ay~e hurt
(ACC)
me.
b. ?Ben (Ane tarafmdan) kmldtm.
Experiencer
Theme
(NOM)
I was hurt by
Ay~e.
c. Ben Ayse'ye kmldtm.
Experiencer
Theme
(NOM)
(OAT)
I felt offended because of Ay~e.
(7) a. Ayse beni sarst1.
Theme Experiencer
(NOM)
(ACC)
Ane gave me a shock.
b. ?Ben (Ayse tarafmdan) sarstldtm.
Experiencer
Theme
(NOM)
I was shocked by
Ay~e.
c. *Ben Ayse'ye/ Ayse'den sarstldtm.
Experiencer
(NOM)
Theme
(DATI ABL)
(8) a. Ayse beni iizdti.
Theme Experiencer
(NOM) (ACC)
Ay~e depressed me.
b. ?Ben (Ay§e tarafmdan) Uztildtim.
Experiencer
Theme
(NOM)
I was depressed by
Ay~e.
155
c. Ben Ay$e'ye tiztildtim.
Experiencer Theme
(NOM)
(OAT)
I felt sorry for Ay~e.
(1)- (8) indicates that passive interpretations do not sound as well as reflexive
ones (without the parenthesized by phrases).
Sentences from (1a) to (8a) imply the direct causation inherent in the meaning
of Type 2 verbs. In these sentences, the Theme does something to make the Experiencer
experience the emotional/ mental state described by the verb. However in (b) sentences:
•
the argument indicated with by phrase is also a kind of Causer of the event
•
but this argument does not actively participate in the event
•
this argument is just a Stimulus for the Experiencer
•
therefore, there is a reflexive meaning rather than a passive one
•
thus, (b )s do not have a high degree of acceptability
Some of the Type 2 verbs may be used as Type 3a verbs as in (3c), (5c), (6c)
and (8c) and some others may be used as Type 3b verbs as in (2c) when the passive
morpheme is attached. Such verbs as buyule-, bog- and sars- can not have dative or
ablative arguments in their passivized forms as the ungrammaticality of(1c), (4c) and (7c)
shows.
Only the ablative (2c) is the exact paraphrase of (2b). This may be the result of
the ablative case marking of the by phrase in Turkish. (3b) and (3c), (5b) and (5c), and,
(6b) and (6c) are not exact paraphrases of each other. As the English translations indicate,
(b) sentences are passive but (c) sentences are not. Therefore, all these verbs have reflexive
meanings not passive.
156
Type 3a Psych Verbs and Passivization
II. 3. 3. 3.
Some verbs of Type 3a take the passive morpheme without any difference in
meaning:
I VERB
. <I MO)lp!JEME
l ay- (to be awaken) -11
1
ay1l- (to be awaken )
reflexive
1
Only the passive morpheme attached forms of some Type 2 verbs belong to
Type 3a:
VERB
·.··.· M;p)lp~ME
-11
-11
-II
-II
-II
/'
bayboz-
kutaktut-
Table 6.
FU;N(J'[JQ~
ii;;/_\ \1.~.
reflexive
reflexive
reflexive
reflexive
reflexive
bay !Ibozul-
kmltakiitutu I-
Active Type 2- Pass1ve Type 3a verbs
Some Type 3a verbs are not separable from their passive morphemes such as
dan/- , and some others have non psych meanings in their passive forms such as yont (to
mean to interpret) and yontul- (to mean to be educated).
Since subjectless sentences are ungrammatical in English (as EPP states), the
Experiencer has to be added to the English translations of (b)s of the following, but in
Turkish, leaving only the dative Theme on the surface is grammatical (lb)- (8b):
(1) a. Ailesi Ay~e'ye gtivendi.
Experiencer
(NOM)
Theme
(OAT)
Her family misted Ay~e.
b.
Ay~e'ye
gtivenildi.
Theme
(OAT)
Ay~e was trusted
by her family.
(2) a. Berkes ona imrendi.
Experiencer
(NOM)
Theme
(OAT)
Everyone was jealous of her.
157
b. Ona imrenildi.
Theme
(DAT)
She was envied by everyone.
(3) a. Halk bu palavralara inandt.
Experiencer
(NOM)
Theme
(DAT)
People believed these lies.
b. Bu palavralara inamld1.
Theme
(DAT)
These lies were believed by the public.
(4) a. Ogrenciler bu zorluklara katland1.
Experiencer
(NOM)
Theme
(DAT)
The students coped with these difficulties.
b. Bu zorluklara katlanlld1.
Theme
(DAT)
These difficulties were coped with by the students.
(5) a. Cocuklar bu odeve yOk ozendi.
Experiencer
(NOM)
Theme
(DAT)
The children took great pains with this homework.
b. Bu odeve yOk ozenildi.
Theme
(DAT)
This homework was taken great pains by the students.
(6) a. Bu cevaba herkes 90k
Experiencer
(NOM)
~a~ud1.
Theme
(DAT)
Everyone was surprised with this answer.
b. Bu cevaba 90k
~a~mld1.
Theme
(DAT)
This answer surprised everyone.
(7) a.
Karde~leri
Experiencer
(NOM)
hastahgma 90k tela~land1.
Theme
(DAT)
Her brothers worried a lot about her illness.
b. Hastahgma 90k tela~lamld1.
Theme
(DAT)
She was worried about a lot because of her illness.
(8) a. Halk ekonomideki
Experiencer
(NOM)
geli~melere
timitlendi.
Theme
(DAT)
The public became hopeful about the economical improvements.
158
b. Ekonomideki geli§melere Umitlenildi.
Theme
(DAT)
It was the economical improvements that people were hopeful about.
II. 3. 3. 4.
Type 3b Psych Verbs and Passivization
The passive morphemes give the following verb a reflexive meaning:
Table 7.
Type 3b verbs with Passive Morphemes with a Reflexive Meaning
One verb of Type 3b; bun a/- can not be separated from its passive morpheme
and there is also one verb of Type 3b which has two different meanings in its -(I)n and -II
forms:
VERB
~ek-
p ASSIYE ' :<.;::,:d::.;.
"to suffer from/ to tolerate"
,~ :-zi~1;t::
·"·~""'to be tolerated"
9ek-il-"to be suffered from/
9ek-in- "to hesitate"
There is a similar case of Type 3b with the Type 3a in terms of the translation
of the passive sentences with deleted Experiencer arguments (which are ungrammatical in
English but acceptable in Turkish):
(1) a. Herkes soguklardan btktt.
Experiencer
(NOM)
Theme
(ABL)
Everyone was fed up with the cold weather.
b. Soguklardan btklldt.
Theme
(NOM)
People were fed up with the cold weather.
(2) a. Cocuklar doktordan korkar.
Experiencer
(NOM)
Theme
(ABL)
The children fear the doctors.
b. Doktordan korkulur.
Theme
(NOM)
People fear the doctors.
159
(3) a. Herkes ondan
ku~kuland1.
Experiencer Theme
(NOM)
(ABL)
Everyone was suspected of him.
b. Ondan ku~kulamld1.
Theme
(ABL)
He was being suspected of.
(4) a. Cocuklar bizden utand1.
Experiencer
Theme
The children were afraid of us.
b. Bizden utanlld1.
Theme
(ABL)
We were being ashamed of.
II. 3. 3. 5.
Type 4 Psych Verbs and Passivization
The following verbs in their psych senses are generally used with the passive
morpheme but they have a reflexive meaning:
VERB
durgerkassiizylrt-
MORPHEME
-Il
-11
-11
-11
-11
Table 8.
FUNCTION
reflexive
reflexive
reflexive
reflexive
reflexive
VERB
durulgerilkasllsiiziily1rtll-
Type 4 Passtve Verbs wtth a Reflextve Meanmg
It is not easy to draw a line between the verbs of Type 4 and the verbs which
have an optional (mostly the dative and sometimes the ablative) argument. Some of the
Type 4 verbs can be used with an optional dative argument. When these verbs are used
together with their dative arguments with the meaning of an implied goal or cause, it is
acceptable to passivize them. In the other case where the verb is used as a Type 4 verb with
only an Experiencer argument, the passivized form is unacceptable.
160
II. 3. 4.
Psych Verbs and Reflexivity in Turkish
Reflexivity in Turkish is a widely discussed phenomenon since Turkish voice
suffixes, can also function as derivational suffixes. This multifunctionality sometimes
causes confusions. The very often confusion is among the passive and the reflexive voice
markers. In some cases phonological conditions determine the attachment of either of
these.
II. 3. 4. 1.
Type 1 Psych Verbs and Reflexivity
Normally the Agent and the Patient arguments of a reflexive verb are
coreferential. Therefore, only the reflexive pronoun kendi can be the object of an
intransitive reflexive verb.
There are four verbs of Type 1 to which the reflexive morpheme -(J)n can be
attached. However, these morphemes do not give a reflexive meaning to the original psych
meaning but rather they produce separate entries which are sometimes psych and
sometimes non psych:
Another instance is the attachment of the reflexive morpheme to the ·Type 1
verb sev- which produces another verb sevin- that belongs to Type 3a:
(1) Ali Ay$e'yi sevdi.
Experiencer
(NOM)
Theme
(ACC)
Ali loved Ay~e.
161
(2) Ali bu habere sevindi.
Experiencer
(NOM)
Theme
(DAT)
Ali is pleased with this news.
Although sevin- seems to be derived from its Type 1 counterpart with the
reflexive -(I)n, it would not be wrong to regard it as a separate entry; sev- with an
accusative animate/animate and sevin- with an inanimate optional dative Theme. Similarly
iste- can be seen as a transitive verb with an accusative animate/animate Theme and istenas an intransitive verb with a unique Experiencer argument.
There are also some psych verbs of Type 1 whose bases seem to be inseparable
from the reflexive morpheme such as kzskan-, ogren- etc.
As a result, ruling out the only seemingly reflexive verbs of Type 1 shows that
there are no originally reflexive psych verbs which belong to Type 1.
II. 3. 4. 2.
Type 2 Psych Verbs and Reflexivity
As the following table exemplifies, some Type 2 verbs allow the reflexive
morpheme to be attached to give a reflexive meaning such as buyule-, etkile- etc. However,
most often, the passive morpheme gives a reflexive meaning:
VERB
bliyUle-
MORPHE
ME
-(l)n
FUNCTION
Ayse
reflexive
Theme
(NOM)
(Ay~e
Ben
Experiencer
(ACC)
charmed me.)
biiyiilendim.
Experiencer
(NOM)
(I was charmed.)
etkile-
-(I)n
reflexive
reflexive
Ayse
Theme
(NOM)
(Ay~e
beni
etkiledi.
Experiencer
(ACC)
attracted me.)
Ben etkilendim.
Experiencer
(NOM)
(I was attracted.)
162
bay-
-Il
passive
reflexive
bog-
-Il
passive
reflexive
boz-
-Il
passive
reflexive
ktr-
-Il
passlVe
reflexive
sars-
-Il
passive
reflexive
liz-
-Il
passtve
reflexive
..
Table 2. Type 2 Verbs and ReflexiVIty
Ben
Ayse'den etkilendim.
Experiencer Theme
(NOM)
(ABL)
(I was attracted by Ane.)
~ beni baydt.
Experiencer
Theme
(NOM)
(ACC)
(Ay~ bored me.)
Ben baytldtm.
Experiencer
(NOM)
(I was bored.)
Ben~ baytldtm.
Experiencer Theme
(NOM) .(OAT)
(I liked Ay~ very much.)
Ayse beni bogdu.
Theme
Experiencer
(NOM)
(ACC)
(Ay~e annoyed me.)
Ben boguldum .
Experiencer
(NOM)
(I was annoyed.)
Ayse beni bozdu.
Theme
Experiencer
(NOM)
(ACC)
(Ay~e embarrassed me.)
Ben bozuldum.
Experiencer
(NOM)
(I was embarrassed.)
~ bozuldum.
Ben
Experiencer Theme
(NOM)
(OAT)
(I felt dis<.:oncerted because of Ay~e . )
Ayse . . beni ~!!~~.. .
Theme Experiencer
(NOM)
(ACC)
(Ay~e hurt me.)
Ben kmldtm.
Experiencer .
(NOM)
(I am hurt.)
Ben
~ ktnldtm.
Experiencer Theme
(NOM)
(OAT)
(I felt offended because of Ay~.)
Ayse beni sarstt.
Theme
Experiencer
(ACC)
(NOM)
(Ay~e gave me a shock.)
Ben sarstldtm.
Experiencer
(NOM)
(I was shocked.)
Ayse beni iizdil.
Experiencer
Theme
(ACC)
(NOM)
(Ay~e distressed me.)
Ben iiziildiim.
Experiencer
(NOM)
(I was disteressed.)
Ben
A~~e·~e iiziildiim.
Experiencer Theme
(NOM)
(OAT)
(I was distressed for Ay~e.)
163
Among these verbs, only the verb etkile- can be used as a Type 3b verb when
the reflexive morpheme is attached. Boz-, kzr- and iiz- can also be used with optional dative
arguments when -II is attached. However -II gives them a reflexive meaning rather than a
passive one. All of the other Type 2 verbs tum to be Type 4 verbs with an only
Experiencer argument in their reflexive form. This may raise the question of whether -(I)n
added to etkile- gives a passive or a reflexive meaning to the verb. (ld) shows that
although the Theme is a kind of Causer, the event is not a passive (1 d) but a reflexive one
(lb/c):
(1) a. Ane
Theme
(NOM)
Ay~e
beni
etkiledi.
Experiencer
(ACC)
attracted me.
b. Ben etkilendim.
Experiencer
(NOM)
I am attracted.
c. Ben
Experiencer
(NOM)
Ay§e'den etkilendim.
Theme
(ABL)
I am attracted to Ay~e.
d. *Ben Ay~e tarafmdan etkilendim.
Experiencer
(NOM)
Causer
*I was attracted by Ay~e.
II. 3. 4. 3.
Type 3a Psych Verbs and Reflexivity
Type 3a verbs which are derived from nouns with -/An have reflexive
meanmgs such as alevlen-, ate~len-, deli/en-, dertlen-, diklen-, duygulan-, ejkarlan-,
heveslen-, huysuzlan-, hiiziinlen-, keyiflen-,
ne~elen-,
ojkelen-, sevdalan-, terslen-, zevklen-
etc. These verbs may have either optional dative Themes or they can be used only with an
Experiencer argument as Type 4 verbs.
164
Other than -/An forms, following forms have also reflexive meanings although
the reflexive morpheme seem to be inseparable from the verb root such as aldan-, giicen-,
giiven-, imren-, inan-, kan-, ozen-, U.Jen-, yuksiin-, zorsun- etc.
II. 3. 4. 4.
Type 3b Psych Verbs and Reflexivity
If we regard -/An as a morpheme which derives intransitive reflexive verbs,
similar to Type 3a, there are also Type 3b verbs of this kind such as endi,Jelen-, esinlen-,
ho#an-,
ku.~kulan-,
onurlan-,
~uphelen-,
umutlan- etc.
Just like verbs of Type 3a, there a few number of verbs of Type 3b which are
inseparable from their seemingly reflexive morphemes such as gocun-, igren-, tiksin-,
usan-, utan-, yrpran- etc.
There is one verb of Type 3b which have different psych meanings with and
without the reflexive morpheme:
VERB
vek-
BARE
bu dertten yok vekti.
-Ali
Experiencer Theme
(ABL)
(NOM)
Ali suffered from this problem a lot.
II. 3. 4. 5.
· REFLEXIVE
'
Ali Ayse'den yok yekindi.
-
Experiencer
Theme
(ABL)
(NOM)
Ali was really ashamed of Ay~e.
Type 4 Psych Verbs and Reflexivity
All of the intransitive verbs of Type 4 derived with -/An have reflexive
meanmgs like aklllan-, bilin9len-, biiyiiklen-, co.Jkulan-, di#en-, gam/an-, heyecanlan-,
hr~rmlan-,
kurt/an- etc.
There is one verb of this type which has different psych meanings with and
without the reflexive morpheme:
165
VERB BARE
kas
Ali kendini bu
i~
REFLEXIVE
Ali 90k kasmd1. (bobtirlendi)
i9in 90k kast1.
Experiencer
(NOM)
Ali forced himself very much for this work.
Experiencer
(NOM)
Ali was very full of self importance.
As the reflexive pronoun indicates, there is a reflexive meaning even in the
case where there is not a reflexive morpheme.
Other than these, there are some verbs of Type 4 like eglen-, tiiken- etc. whose
bases can not be separable from the reflexive morpheme.
As we mentioned above there are a number of verbs of this kind which can not
be separated from their seemingly reflexive morphemes, the etymological reason of which
is not the concern here:
TYPE
Type 1
Type 1
Type 3a
Type 3a
Type 3a
Type 3a
Type 3a
Type 3a
Type3a
Type3a
Type 3a
Type 3a
Type 3b
Type 3b
Type 3b
Type 3b
Type 3b
Type 3b
Type4
Type4
Table 3.
VERB
k1skan-
-
~gren-
aldangiicengiivenimreninankanozenU~en-
yiiksiinzorsungocunigrentiksinus anutany1praneglentiikenSeemmgly Reflex1ve Psych Verbs
166
II. 3. 5.
Psych Verbs and Reciprocity
The data shows that although a few number of psych verbs in the data allow the
reciprocal morpheme to be attached, in general, these verbs do not have a reciprocal
meaning. Semantically this may be due to the reason that psych activities are personal
matters which can not be done reciprocally or cooperatively:
( 1) a. Ali ile Ay!}e birbirlerini ozlediler.
Experiencer
(NOM)
Theme
(ACC)
Ali and Ane missed each other.
b. ?Ali ile Ay!}e
ozle~tiler.
Experiencer
(NOM)
Some psych verbs can be used with the reciprocal morpheme but the meaning
is not reciprocal. Therefore (2c) seems more acceptable than (2b):
(2) a. Ali Aylle'ye k1zd1.
Experiencer Theme
(NOM) (OAT)
b. ?Ali ile Ayse birbirlerine klZl~tllar.
Ali
Experiencer
(NOM)
and Ay~e got
c. Ali Ay$e'ye
Theme
(OAT)
angry with each other.
klZl~ti.
Experiencer Theme
(NOM)
(OAT)
Some other verbs have different meanings with and without the reciprocal:
(3) a. Ali ile Ay!}e birbirlerini anladllar.
Experiencer
(NOM)
Ali and Ay~e
Theme
(ACC)
understood each other.
b. Ali ile Ay!}e
Ali
Experiencer
(NOM)
and Ay~e got
anla~tllar.
along well with each other.
(4) a. Ali ile Ay!}e birbirlerini seviyorlar.
Experiencer
(NOM)
Theme
(ACC)
Ali and Ane love each other.
167
b. Ali ile Ayse
sevi~iyorlar.
Experiencer
(NOM)
Ali and Ane are making love.
There are some reciprocal psych forms in the Old Turkish like sav- i~'- to mean
to like to love one another.
In all of the examples, the reciprocal pronoun birbiri is the Theme argument
and the two participants are regarded as one Experiencer argument. Regarding them as
separate Experiencers is ungrammatical. Kural's reason for the absence of ExperiencerExperiencer verbs is that the Experiencer subject makes the verb stative but Experiencer
object forces the verb to be eventive and the result is a clash between the two.
Moreover, if we regard them as separate arguments, it causes the violation of
the Theta Criterion:
anla~tllar.
Experiencer Experiencer
(5) *Ali ile Ane
(NOM)
(NOM)
Ali and Ane underdstood each other.
II. 4.
Compound Psych Uses in Turkish
There are two ways of expressing psych events in Turkish; first by lexical
means and second by compound constructions. There are 31 helping verbs which are the
components of these compound constructions; a/zn-, a/-, at-, bag/a-, besle-,
davran-, don-, dur-, duy-,
du~-.
bo~alt-,
r;ek-,
edin-, et-, ger;ir-, gel-, getir-, git-, gar-, gaster-, kal-, kapzl-,
kesil-, ol-, tut-, tutu/-, ugra, uyandzr-, var-, ver-, yap-.
These helping verbs can either be combined with an adjective, a noun or a
complex nucleus.
168
II. 4. 1.
Compound Psych Verbs which Derive with a Nominative Psych
Nominal
Most of the compound constructions are combined with a nominative psych
nominal and a helping verb.
b~gla-
A psych nominal together with the helping verb
produces a Type 3a
compound psych verb:
NOMINAIJ
CASE
·HELPING
VERB
SENSE · ·
iimit
nominative
ba~la-
psych
.
•
v
TYPE ...
...... ·...•
~'
Type 3a
( 1) Ali bu i§e limit baglad1.
Experiencer Theme
(NOM)
(OAT)
Ali hopes for this job.
A psych nominal with besle produces a Type 3a psych verb too:
(2) Ay;;e Ali 'ye derin bir sevgi besliyor.
Experiencer Theme
(NOM) (OAT)
Ay~e loves Ali
deeply.
.~Jdf!.NQ
~R.B ··
nefret
sevgi
umut
iimit
Table 2.
nominative
nominative
nominative
nominative
beslebeslebeslebesle-
psych
psych
psych
psych
Type
Type
Type
Type
3a
3a
3a
3a
Compound Psych Verbs Which Derive with a Nominative Psych Nommal (2)
The helping verb
9ek~
can be compounded with a nominative nominal
becoming a Type 4 psych verb which can be used with an optional purpose clause:
(3) a. Ali ryok act ryekiyor.
Experiencer
(NOM)
Ali feels great pain.
169
b. Ali Ayse icin c;ok act c;ekti.
Experiencer Theme
(NOM)
(PURP.)
Ali felt great pains for Ay~e.
The most productive helping verb which can be compounded with almost all
psych nouns is duy-. There is not a significant meaning difference between the psych
compounds with duy- and their lexical counterparts:
NOMINAL
· CME
aCI
arzu
gurur
heyecan
ilgi
istek
ozlem
utan9
nominative
nominative
nominative
nominative
nominative
nominative
nominative
nominative
HEI.iP,;JN(;
SENSE
TYPE
duyduyduyduyduyduyduyduy-
psych
psych
psych
psych
psych
psych
psych
psych
Type4
Type 3a
Type 3b
Type4
Type 3a
Type 3a
Type 3a
Type 3b
viRti
Compound Psych Verbs Whtch Denve wtth a Nommattve Psych Nommal (4)
Table 4.
(4) Ayse c;ok heyecan duydu/ heyecanlandt.
Experiencer
(NOM)
Ali felt a great excitement.
(5) Ayse dilbilime ilgi duyuyor/ ile ilgileniyor.
Experiencer
(NOM)
Ay~e
(6)
Theme
(DAT)
is interested in linguistics.
!sJ.3.e soylediklerinden utanc; duyuyor/ utamyor.
Experiencer
(NOM)
Ay~e
Theme
(ABL)
feels ashamed for what she says.
The helping verb et- produces four different types of compound psych verbs
either from bare nominals:
170
or from the ones derived with -llk:
NOMINAL
'
'
CASE
aksilik
arstzhk
cadlltk
.-.:.
. HELPING
,,pRB
· ·. .
;
densizlik
mizmizhk
.,..
.;·~.
nominative
nominative
nominative
nominative
nominative
nominative
~ocukluk
Table 6.
".
'~
etetetetetet-
. ,.
"
SEN~E
. .. .
'
,-;, '
psych
psych
psych
psych
psych
psych
l£5L .
.
~
_:::
Type 4
Type 4
Type 4
Type 4
Type 4
Type 4
. •.
__;,___
Compound Psych Verbs Which Derive with a Nominative Psych Nominal (6)
(7) Ali bu giinleri umut etti.
Experiencer
(NOM)
Theme
(ACC)
Ali hoped for these days.
(8) Cocuklar dondurmaya heves etti.
Experiencer
(NOM)
Theme
(DAT)
The children had a fancy for the ice cream.
(9) Doktorlar kanamasmdan endi~e etti.
Experiencer
(NOM)
Theme
(ABL)
The doctors worried about her bleeding.
( 10)
Cocuklar kilosuyla alay etti.
Experiencer
(NOM)
Theme
(INST)
The children made fun of her weight.
(11)
Bebek m1zn11zhk ediyor.
Experiencer
(NOM)
The baby is grousing.
The following psych nominals with ge9ir- derive Type 4 compound psych
verbs:
( 11)
Kadm sonunda cinnet ge9irmi~.
Experiencer
(NOM)
The woman got out of her mind at the end.
171
There are some compound psych verbs with gel-+ nominative psych nouns.
This combination produces Type 4 verbs however; the Experiencer argument of these
verbs is not nominative but dative:
(12)
Bana fenahk geldi.
Experiencer
(OAT)
I feel bored/ fainted.
These verbs show that our classification of psych verbs in Turkish should be
revised to include the compound uses. Therefore, there is. another class which we will call
as Type 4a. Type 4a is valid only for a group of compound verbs. These verbs have only an
Experiencer argument just like Type 4 verbs but different from them, their Expeliencer is
not nominative but dative marked just like the dative argument ofType 3a verbs:
TYPE
SUBJECT
OBJECT
4a
Experiencer (DAT)
0
Gaster- can also be combined with a psych noun but this time a different
situation emerges as to the type of the compound psych verb. In these cases the dative
marked object argument is not a Theme but an Experiencer and the nominative marked
subject is a Theme. Therefore, another type should be added to our classification of psych
verbs. This type is also valid for a number of compound uses whose case marking is just
like Type 3a verbs whereas the order of the Theme and the Experiencer is just like Type 2
verbs. This will be referred as Type 2a verbs:
TYPE
2a
SUBJECT
.OBJECT
Theme(NOM)
Experiencer (DA T)
172
nominative
nominative
sevgi
yakmhk
Table 9.
gostergoster-
psych
psych
Type 2a
Type 2a
Compound Psych Verbs Which Derive with a Nominative Psych Nominal (9)
Kadm bu cocuklara yeterince sevgi gostermemi~.
(13)
Theme
(NOM)
Experiencer
(OAT)
The woman did not love these children enough.
(14)
Kom$ular bize yok yakmhk gosterdi.
Theme Experiencer
(NOM)
(OAT)
The neighbors showed us a great hospitality.
Similarly, ver- with a psych noun sometimes produces Type 3a verbs in terms
of dative case marking while they are Type 2 verbs in terms of the position of the
Experiencer argument. Although they are Type 2a verbs, some other compound psych ·
verbs with ver- are Type 3a verbs:
NOMINAL ··
~.A~F;.
i
btkkmhk
daral
heyecan
onem
tepki
'
nominative
nominative
nominative
nominative
nominative
HE.L~IN9 .•.
VERB.·
ververververver-
·.
t;
1$F,;JY~~
·, · ·· . TYPE '
psych
psych
psych
psych
psych
.
Type 2a
Type 2a
Type 2a
Type 3a
Type 3a
Table 10. Compound Psych Verbs Which Derive with a Nominative Psych Nominal (10)
(15)
Ali Ay$e'ye daral/ heyecan/ btkkmhk verdi.
Theme Experiencer
(NOM) (OAT)
Ali bored
(16)
Ay~e.
Ali Ay§e'ye onem verdi.
Experiencer
(NOM)
Theme
(OAT)
Ali cared about Ay~e.
A psych nominal with uyandzr- produces a similar case which we will call as
Type 2b psych verbs such as the following ones which have locative Experiencers and
nominative Themes just like the order in Type 2 as (17a) and (18a) shows. Uyandzr- can
also be used without the causative morpheme appearing either with a dative Theme or a
purpose clause as in (17b), (17c) and (18b), (18c):
173
(17)
a. Bu tip konular bende a~m ilgi uyandmyor.
Theme
(NOM)
Experiencer
(LOC)
This kind of subjects interest me a lot.
b. Bu tip konulara (kar~1) bende ~m bir ilgi uyantyor.
Theme
(OAT)
Experiencer
(LOC)
c. Bu tip konular igin bende a~m bir ilgi uyan1yor.
PURP ~
(18)
CL.
Experiencer
(LOC)
a. Sokak. s:ocuklan kiminde merhamet kiminde nefret uyandmyor.
Theme
(NOM)
Experiencer
(LOC)
Homeless children make some feel pity and some others feel hate.
b. Sokak s:ocuklarma kar~1 insanlarda merhamet uyamyor.
Theme
(OAT)
Ex.periencer
(LOC)
c. Sokak. c;ocuklan ic;in insanlarda merhamet uyan1yor.
PURP. CL.
Ex.periencer
(LOC)
The helping verb yap- can be combined with either bare psych nouns:
or with nouns derived with -Ilk. When it is used with a psych noun derived
with -Ilk, it produces Type 4 psych verbs:
NOMINAL
CASE
HELPING
VERB
·SENSE
ahmakhk
nominative
nominative
yapyap-
psych
psych
~apkmhk
Type4
Type4
~ocukluk
serserilik
sululuk
~tmankhk
terbiyesizlik
yapyapyapyapyap-
nominative
nominative
nominative
nominative
nominative
psych
psych
psych
psych
psych
Type4
Type4
Type4
Type4
Type4
Table 13. Compound Psych Verbs Whtch Denve wtth a Nommattve Psych Nommal (13)
(19)
Ktz hazen 90k ~tmankhk yaptyor.
Experiencer
(NOM)
The girl sometimes spoils a lot.
However when it is used with bare psych nominals it produces either a Type 3a
as in (20) or a Type 2a psych verb as in (21):
(20)
Ali Ayse'ye cilve yaptyor.
Experiencer Theme
(NOM)
(OAT)
· Ali flirts with Ay~e.
(21)
Ali Ayse'ye baskt yaptyor.
Theme Experiencer
(NOM) (OAT)
Ali pressures on
Ay~e.
The analyses in this section have shown that most of the compound psych
verbs in Turkish are constructed by combining a nominative psych nominal with such
helping verbs as bag/a-, besle-, f;ek-, duy-, et-, gef;ir-, gaster-, uyand1r-, ver- and yap-:
The output compound psych verb can belong to one of the specified psych verb
classes; Type 1, 3a, 3b, 4 or to three new classes; 2a, 2b and 4a (as variants of Type 2 and
4) which should be added to the existing four psych verb types.
175
II. 4. 2.
Compound Psych Verbs which Derive with an Accusative Psych
Nominal
Compared to the compound psych verbs derived with nominative nominals,
accusative ones are lesser in number. For example
~ek-
can be used with a genitive+
accusative marked psych noun similar to-a Type 1 psych verb:
NOMINAL
CASE
HELPING
SENSE
lfYPE
actsmt
hasretini
ozlernini
accusative
accusative
accusative
~ek-
psych
psych
psych
Type 1
Type 1
Type 1
VEfia
~ek-
~ekTable 14. Compound Psych Verbs Wh1ch Denve With an Accusai:Jve Psych Nommal (I)
(1) Ali ytllarca Ay~e'nin hasretini ~ekti.
Experiencer
(NOM)
Theme
(ACC)
Ali yearned for Ay~e for long years.
Without genitive marking, these helping verbs can be used with .nominative
psych nominals as acz
~ek-,
ozlem
~ek-,
hasret 9ek- etc.
The following genitive+ accusative psych constructions are used with yerine;
they become Type 4 compound psych verbs with gel and they can also be used with getir(which is inherently causative) together withyerine as in (3):
(2) Ali 'nin keyfi yerine geldi.
Theme
(NOM)
Ali cheered again.
(3) Ay§e Ali'nin keyfini yerine getirdi.
Causer
Theme
(NOM)
(ACC)
Ay~e made Ali cheer
again.
NOMINAL
CASE
HELPING
SENSE
VERB
keyfi(ni)/ ne~esi(ni) yerine
dative
gel-/ getir-
psych
Table 15 . Compound Psych Verbs Wh1ch Denve w1th an Accusative Psych Nommal (2)
176
Kes- can also be used with an accusative psych noun as a Type 3b verb:
(4) Ali Ayse'den funidi kesti.
Experiencer
The~
(NOM)
(ABL)
Ali gave up his hopes for
Ay~e.
The following accusative nominals can be combined with the verb tut- which
produces the following expressions that require a genitive Experiencer. This is a variant of
Type 4 which will be called as Type 4b thereafter:
I
NOMINAL
~tocuklugu
iyiligi
SUBJECT . ·. . ·· .
I~BJECT
CASE
HELPING
SENSE
TYPE
accusative
accusative
accusative
VERB
tuttuttut-
psych
psych
psych
Type4
Type4
Type4
';.
•'
li·.;:.·
yardtmseverligi
Table 17. Compound Psych Verbs Wh1ch Denve w1th an Accusative Psych Nommal (4)
(5) Ali'nin iyiligi tuttu.
Experiencer
(GEN)
Ali became such a kind hearted person.
Therefore, there are only three helping verbs; rek-, kes-, gel-, getir- and tutwhich can be used with an accusative nominal to produce a compound -psych verb. The
output verbs belong to Type 1, Type 3b or a proposed new class called Type 4b.
177
II. 4. 3.
Compound Psych Verbs which Derive with a Dative
Psych Nominal
AI- as a helping verb produces non psych compound constructions when
combined with a nominative or accusative nominal. When it is used with a dative psych
nominal it either produces Type 1 or Type 2 compound psych verbs:
hedef
onlem
nasibini
alaya
iistiine
ciddiye
nominative
nominative
accusative
dative
dative
dative
alalalalalal-
non psych
non psych
non psych
psych
psych
psych
Type 2
Type 1
Type 1
Table 18. Compound Psych Verbs Which Derive with a Dative Psych Nominal (I)
(1) Ali Ay;;e'yi ciddiye almad1.
Experiencer Theme
(NOM)
(ACC)
Ali did not take Ane seriously.
(2) Ali Ayse'yi alaya aldt.
Theme Experiencer
(NOM) (ACC)
Ali made fun of Ay~e.
AI- can produce another Type 1 verb when the reflexive morpheme is attached
to the verb:
·HELPING ·
· :VER!J.·_··.·. ..
Table 19. Compound Psych Verbs Which Derive with a Dative Psych Nominal (2)
(3) Ayse herseyi Ustline ahruyor.
Experiencer Theme
(NOM)
(ACC)
Ay~e
takes everything as offence.
The compound of dative psych nouns with dii,J- results in Type 4 compound
psych verbs:
NOMII'l'AL
··-·-·- --- ---· -.. ---
a~ka
CASE
dative
dative
HELPING
-...Yi.R:B' . .dii~-
SENSE · '
-- ,_ -·.
psych
psych
TYPE
- "' ---~ . .•---·----·-·
Type 4
dli~Type 4
Table 20. Compound Psych Verbs Which Derive with a Dative Psych Nominal (3)
bunahma
178
(4)
Ay~e
son zamanlarda iyice bunahma dii~ti.i.
Experiencer
(NOM)
Ay~e has
been in a real depression recently.
There are other instances where the Experiencer is in the object position. The
difference from the Type 2 is in the case marking. When the following dative psych
nominals are used with gel, the Experiencer is either a genitive marked nominal (Sa) or an
overt possessive adjective. This type of compound psych verbs will be named as Type 2c
verbs which are another instantiation of Type 2 verbs:
TYPE
2c
~a~kmltgma
dative
gel-/ getirpsych
TYJ>e 2c
Table 21. Compound Psych Verbs Which Derive with a Dative Psych Nominal (4)
~a~kmhg1ma
(5) a. Sordugu soru
geldi.
Theme
(NOM)
I was in a rather dumfounded mood when she asked the question.
b. Sordugu soru benim ~a~kmhg1ma geldi.
Theme
(NOM)
Experiencer
(GEN)
With getir- which is causative itself, there is a Causer, an accusative Theme and
again a genitive Experiencer:
(6) Ali bu soruyu ~a~kmhg1ma getirdi.
Causer
(NOM)
Theme
(ACC)
Ali deliberately asked this question when I was in such a dumfounded mood.
The following genitive+ dative psych nouns with git- produce Type 2c psych
verbs too:
giiciine
ho~una
tuhafma
dative
dative
dative
gitgitgit-
psych
psych
psych
Type 2c
Type 2c
Type 2c
Table 22. Compound Psych Verbs Which Derive with a Dative Psych Nominal (5)
179
(7) Ali'nin sozleri Ay~e'nin ho~una gitti.
Theme
(NOM)
Ay~e liked
Experiencer
(GEN)
what Ali said.
Kapzl- and ugra- with dative psych nouns produce Type 4 compound psych
verbs:
(8)
Ay~e
birden tela~a kaptldt.
Experiencer
(NOM)
Ane suddenly panicked.
HELPING
NOMINAL'
\TER.B
~a~kmhga
dative
ugrapsych
Type 4
Table 24. Compound Psych Verbs Which Derive with a Dative Psych Nominal (7)
(9)
~bu
haberi duyunca dumura ugradt.
Experiencer
(NOM)
Ay~e
became stupefied when heard the news.
The following compound psych .verbs with var- require genitive marked
Themes. This presents another instance which will be regarded as Type 5 that does not
correspond any of the proposed types:
NOMINAL.
Table 25. Compound Psych Verbs Which Derive with a Dative Psych Nominal (8)
180
(10)
Ali gerceklerin farkma vard1.
Experiencer
(NOM)
Theme
(GEN)
Ali realized the truths.
(11)
Ali bu olaylann aynmma vard1.
Experiencer
(NOM)
Theme
(GEN)
Ali figured out these events.
II. 4. 4.
Compound Psych Verbs which Derive with a Complex Nucleus
The following compound psych verbs consist of two verbs one of which is a
psych verb derived with -Ip and the other is one of the helping verbs at-, git- and kal-.
These helping verbs reinforce the meaning of the psych component:
dahp
unutup
gitgitkatkalkal-
apt~tp
donup
~a~mp
psych
psych
psych
psych
psych
Type 3a
Type l
Type 4
Type 4
Type 3a
Table 26. Compound Psych Verbs Which Derive with a Complex Nucleus (I)
(1)
Ay~e
eski gtinleri unutup gitti.
Experiencer
(NOM)
Ay~e
(2)
has already forgotten the old days.
Ay~e
uzaklara dahp gitti.
Experiencer
(NOM)
Ay~e
Theme
(ACC)
Theme
(OAT)
become lost in thoughts looking far away.
(3) Ay§e donup kald1.
Experiencer
(NOM)
Ay~e
II. 4. 5.
was paralyzed.
Compound Psych Verbs which Derive with an Adjective
The helping verb can be combined with a psych adjective as well as a psych
noun. The resulting compound verb can be a Type 1 verb as in the following:
181
biiyiik
gorgorgortut-
hor
kiiyiik
yeg
psych
psych
psych
psych
Type
Type
Type
Type
1
1
1
1
Table 27. Compound Psych Vetbs which Derive wtth an Adjective (I)
(1) Ali kendini (herkesten)bUyUk gorliyor.
Experiencer
(NOM)
Ali is full of self conceit.
(2) Ali Ay§e'yi hor gorliyor.
Experiencer Theme
(NOM)
(ACC)
Ali looks down upon
Ay~e.
or it can be a Type 2 verb with an accusative Experiencer:
ADJECTIVE
alakadar
mutlu
HELPING
VERB
etet-
SENSE · .'·
TyPE· •
psych
psych
Type2
Type2
Table 28. CqmpoundPsych Verbs whtch Denve wtth an AdJeCtive (2)
(3) Bu konu beni alakadar ediyor.
Theme Experiencer
(NOM) (ACC)
This subject interests me.
(4) Ay§e Ali'yi mutluediyor.
Theme Experiencer
(NOM)
(ACC)
Ay~e makes Ali happy.
The following are Type 3a compound psych verbs derived with adjectives:
(5) Ali Ay§e'ye a~1k oldu.
Experiencer Theme
(NOM)
(DAT)
There are also compound psych verbs which belong to Type 3b in which the
ablative expresses the cause/ reason of the psych event:
182
(6) Ayse otkeden deliye don<;lii.
Experiencer Theme
(NOM)
(ABL)
Ay~e
got mad wih anger.
(7) Ayse herseyden mutlu olur.
Experiencer
(NOM)
Theme
(ABL)
Everything makes Ay~e happy.
(8) Ayse bu sesten tedirgin oldu.
Experiencer
(NOM)
Theme
(ABL)
This sound made Ay~e restless.
Lastly, there are Type 4 compound psych verbs derived with adjectives:
ADJECTIVE
.
9ekingen
utangac;
aktlh
uslu
bozum
sa mimi
tuhaf
HELPING
···
,;- . .. _,._
:VERB
davrandavrandurdurololol__ ..
;
...
>
SENSE
.
',. ~;
' •
psych
psych
psych
psych
psych
psych
psych
1 ·~~~
... .;'i:l~ Z!\t';.:
Type4
Type4
Type4
Type4
Type4
Type4
Type4
Table 31. Compound Psych Verbs whtch Denve wtth an Adjective (5)
(9) Ktz utangay davrandt.
Experiencer
(NOM)
The girl behaved shamefully.
(10)
Cocuklar uslu durdu.
Experiencer
(NOM)
The children stayed quiet.
(11)
Ayse bozum oldu.
Experiencer
(NOM)
Ay~e
was embarrassed.
The analysis of compound .psych verbs has shown that the proposed
classification for lexical psych verbs in Turkish does not cover compound configurations
183
properly. Therefore the classification mles for psych verb types in Turkish should be
rewritten for compound psych verbs adding the six types; Type 2a, 2b, 2c, 4a, 4b and 5:
1
2
Experiencer (NOM)
Theme (NOM)
Theme (ACC)
Experiencer (ACC)
Lexical
Lexical
Theme (NOM)
Experiencer (NOM)
Experiencer (NOM)
Experiencer (NOM)
Experiencer (GEN)
Theme (DA T)
Theme (ABL)
0
CompoUnd
Lexical
Lexical
Lexical
2a
2b __
2c
3a
3b
4
4a
4b
5
E[tperiertcer (GEN) ·
Experiencer (NOM)
· 0
_
.-. . .
Theme (GEN) •·
'
' ·compouri.d
'Compound
Table 32. A New Classification for Turkish Psych Verbs Including Compound Forms
Compound psych verbs can be regarded as a result of an incorporation process.
When the thematic arguments of non psych senses of these helping verbs are examined, it
is observed that there are two types of incorporation; one is Theme Incorporation and the
other is Goal Incorporation as the following table exemplifies:
1
Goal
Incorporation
NOMINAL
CASE
HELPING
VERB
TYPE
alaya
dative
al
Type 2
PSYCH
NON PSYCH
Ali Ayse'yi eve ald1.
Agent Theme Goal
(NOM) (ACC) (DA T)
ciddiye/iistiine
dative
al
Ali
Ayse'yi alaya aldt.
Theme Experiencer
(NOM) (ACC)
Ali
Type l
Ayse'yi ciddiye ald1.
Experiencer Theme
(ACC)
(NOM)
iistiine
2
iimit
dative
nominative
aim
bagla
Type 3a
Ali agaca ffi baglad1.
Agent Goal Theme
(NOM) (OAT) (NOM)
Theme
Incorporation
3
Theme
Incorporation
-
Type 1
sevgi!
nefret/
umut/iimit
nominative
besle
Type 3a
Ali kedi besliyor.
herseyi Ustiine ahmyor.
Ali
Experiencer Theme
(NOM)
(ACC)
Ali
bu ise Umit bagladt.
Experiencer Theme
(NOM) (OAT)
Ali Ayse'ye derin bir sevgi besliyor.
Experiencer Theme
(NOM) (OAT)
Agent Theme
(NOM) (NOM)
*Ali Ayse'ye kedi besliyor.
Agent Goal
Theme
(NOM) (OAT) (NOM)
4
Theme
Incorporation
act! i~kence
nominative
c;ek
Type4
Ali
kura c;ekiyor.
Agent Theme
(NOM) (NOM)
hasretini/ actsmt!
ozlemini
accusative
5
Goal
Incorporation
bunahma/ a~ka
dative
6
Theme
Incorporation
act/ heyecan
c;ek
Type 1/5
Ali
ill! c;ekiyor.
Type4
Ali
yere dU~tU.
duy
Type4
Ali bir ses duyuyor.
nominative
utanc;/ gurur
nominative
duy
Ali bunahma dU~tU.
Ali c;ok act duyuyor.
Experiencer
(NOM)
Agent Theme
(NOM) (NOM)
ilgi! arzu/ istek/
ozlem
Ayse'nin hasretini c;ekiyor.
Experiencer
(NOM)
Patient Goal
(NOM) (OAT)
nominative
Ali
act c;ekiyor.
Experiencer
Theme
(POSS.)
(NOM)
(GEN)
(ACC)
Agent Theme
(NOM)(ACC)
dU~
Ali ~ok
Experiencer
(NOM)
Ali anlambilime ilgi duyuyor.
Type 3a
Experiencer Theme
(OAT)
(NOM)
duy
Type3b
Ali yan odadan sesler duyuyor.
Agent
(NOM)
Source
(ABL)
Theme
(NOM)
Ali soylediklerinden utanc duyuyor.
Experiencer
(NOM)
Theme
(ABL)
X
~.
a lay
7
nominative
et
-
Type 3d
Ali
Ayse'yle alay etti.
Experiencer Theme
(NOM)
(INST.)
endi~e
nominative
et
Type3b
Doktorlar kanamasmdan
Experiencer
(NOM)
aldm~/
heves/
cUret/ te Ia~
nominative
et
Type 3a
nominative
et
etti.
Cocuklar dondurmaya heves etti.
Experiencer
(NOM)
umut/ arzu/ limit
endi~e
Theme
(ABL)
Type 1
Ali
Theme
(ABL)
bu giinleri umut etti.
Experiencer Theme
(NOM)
(ACC)
ak!lhhkl aksilikl
ars1zhk/ cadihk/
((Ocuklukl
densizlik/
mizmizhk
nominative
8
Theme
Incorporation
cinnet/
bunahm
nominative
9
fenaltkl sinir/ bay
buhran
Type4
Ali ak1lhhk ediyor.
Experiencer
(NOM}
ge9ir
Type4
Kadm kotti bir hastahk ge9irmi~.
Patient
(NOM)
dalgmhgma!
Goal
Incorporation
et
nominative
dative
gel
gel/ getir
Type 4a
Type 2c
Kadm sonunda cinnet ge9irmi~.
Theme
(NOM)
Experiencer
(NOM)
Bize misafrr geldi
Bize fenahk geldi.
Goal Agent
(OAT) (NOM)
Experiencer
(OAT)
Ane evime geldi.
Sordugu soru Ayse'nin dalgmh!tma
Agent Goal
(NOM)(OAT)
~a~kmhgma
Theme
(NOM)
geldi.
10
Goal
Incorporation
garibine/ giiciine/
ho~una! tuhafma
11
Theme
Incorporation
ilgi/
yakmhk
sevgi/
12
Goal
I!tcorporation
heyecana!
iimide
tela~a!
dative
git
Type 2c
Ali Ayse'nin okuluna gitti.
Agent
(NOM)
nominative
goster
Type 2a
Ali'nin sozleri Ayse'nin
Goal
(OAT)
Koml!ular bize
Theme
(ACC)
ev/ evi gosterdi.
-~
~--~~-
kapll
Type4
~
ho~una
gitti.
Experiencer
(NOM)
Komsular bize fYOk yakmhk gl>sterdi.
Agent Beneficiary Theme
(NOM) (OAT) (NOM/ ACC)
dative
Experiencer
(GEN)
(POSS)
(OAT)
Experiencer Theme
(NOM)
(OAT)
akmt1ya kapildi.
~ birden tela~a
Experiencer
(NOM)
Agent Goal
(NOM) (OAT)
-
kapiidi.
X
~:
accusative
kes
13
Theme
Incorporation
iimidi
14
Theme
Incorporation
iyiligi/ fYOCUkJugu/
yard1mseverligi
accusative
tut
Type4b
15
ilgil
merhamet/
nefret/
sevgil
nominative
uyandtr
Type 2b
Type 3b
Ali ma~tmtzdan bu parayt kesti.
Agent Source
Theme
(NOM) (ABL) (ACC)
Ali Ayse'den iimidi(n)i kesti.
Experiencer Theme
(NOM) (ABL)
Ali 'nin dilegi tuttu.
Agent Theme
(GEN) (ACC)
Ali'nin iyiligi tuttu.
Experiencer
(GEN)
-
~evkat
farkmal aynmma
dative
var
Type 5
Ali Ankara 'ya vard1.
Agent Goal
(NOM) (OAT)
Ali gerceklerin farkma vard1.
Experiencer Theme
(NOM)
(GEN)
17
Theme
Incorporation
btkkmhkl
heyecan
nominative
ver
Type 2a
Ali Ayse'ye para verdi.
Agent Goal Theme
(NOM) (OAT) (NOM)
nominative
ver
Type 3a
Ali Ayse'ye b1kkmhk verdi.
Theme Expetiencer
(NOM) (OAT)
Ali Ayse'ye onem verdi.
Experiencer Theme
(NOM) (OAT)
18
Theme
Incorporation
onernl tepki
bask!
nominative
yap
Type 2a
cilve/ naz
nominative
yap
Type 3a
ahmakhk/
yocuklukl sululukl
nominative
yap
Type4
~imankhkl
Ayse Ali'ye kek yap1yor.
Agent Beneficiary Theme
(NOM) (OAT) (NOM)
Ayse fYOk giizel kek yaptyor.
Agent
Theme
(NOM)
(NOM)
terbiyesizlik
rable 2. Summary of the Psych and Non-Psych Compound forms
==
Bu tip konular bende a~m ilgi
Theme Experiencer
(LOC)
(NOM)
uyandmyor.
16
Theme
Incorporation
darall
><
><
Ali Ayse'ye bask1 yap1yor.
Theme Experiencer
(NOM) (OAT)
Ali Ayse'ye cilve yap1yor.
Experiencer Theme
(NOM)
(OAT)
~ bazen ~1mankhk yap1yor.
Experiencer
(NOM)
I
I
184
II. 5.
Psych Adjective Derivation
This section analyzes the adjectives which have psychological senses. The
analysis shows that morphologically, psychological state adjectives can be classified into
five groups according to their roots which enter as the input of the adjectivalization
process. These are:
1. those which have verbal roots
2. those which have adjectival roots
3. those which have nominal roots
4. those which are non-derived in any way
5. those which are used as compound forms
On the other hand, semantically, there are two participants m the event
described by the psychological verb. First, there is a Stimulus which is a simple animate or
inanimate NP, an act or a psychological event which stimulates the Experiencer to
experience the psychological state. Second, there is an animate Experiencer who
experiences the mentioned psychological state:
Experiencer
Psychological State Verb
•
•
Stimulus
•
Figure I . The Experiencer Stimulus Interaction
The derivation of psychological adjectives is assumed to include the following
stages:
•
The Stimulus stimulates the psychological state in the Experiencer.
•
The Experiencer experiences the psychological state stimulated by the
Stimulus.
As a result of this bidirectional interrelationship:
185
•
The Experiencer gains some qualities through the effect of the Stimulus
and these qualities are expressed by a modifying psych adjective.
•
The Stimulus gains some qualities through the effect of the Experiencer
and these qualities are expressed by a modifying psych adjective.
Therefore, semantically, there are two main types of psych adjectives to which
a third one which have the both functions could be added:
1. those which modify the Experiencer argument
2. those which modify the Stimulus argument ·
J. those which modify both the Experiencer and the Stimulus arguments
II. 5. 1.
Psych Adjectives Derived from Verbal Roots
In this. group derived from verbal roots, first there are psych adjectives which
modify the Experiencer argument.
-GAn is one of the morphemes which derive Experiencer modifying psych
adjectives from the psych verbs:
1NPUTVERB.
aimazbaybezbiltyekindaldiren-igrenkas1lkmlk1zsiiziiz-
yetinyii
"'
MORPHEME
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
-GAn
Table I,
;
PSYCH AJ)JECTIV:F;
almgan
azgm
baygm
bezgin
bilgin
tyekingen
kalgm
direngen
igrengen
kas1lgan
kmlgan
k1zgm
siizgiin
iizgiin
yetingen
y1lgm
Psych Adjectives Denved Wlth -Gan
186
-GI + cll III slz and -glr; also derive psych adjectives which modify the
Experiencer:
INPUT~RB
bil-
-GI + clllll slz
duy-
-GI + ell ill slz
say-
-GI + ell II/ slz
sevseziiz-
-GI + ell II/ slz
-GI +ell ll/ slz
-GI + cl/ll/ slz
Table 2.
INPUT VERB
bilgici
bilgili
bilgisiz
duygulu
duygusuz
sayg1h
saygiSIZ
sevgili
sezgili
iizgiilii
Psych Adjectives Derived with -GI + cl/111 slz
MORPHEME
PSYCH ADJEC':{IVE
bil
Table 3.
Psych Adjectives Derived with
-gl~
-Ik is one of the most productive morphemes which derive psych adjectives
from psych verbs. All of the following adjectives modify the Experiencer:
INPUT VERB
MORPHEME
PSYCH ADJECTIVE
dongilcenkopkudur-
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-Ik
-lk
donuk
giicenik
kopuk
kuduruk
~Imar-
s1ma~-
U9uyanuyu~-
yih~yirt~
YIVI~-
Table 4.
~1mank
sima~Ik
u~uk
uyamk
uyu~uk
yih~Ik
yutik
YIVI~Ik
Psych Adjectives Denved Wlth -lk
-Inti+ (Ill slz) produces the following adjective forms from the psych verbs:
INPUT VERB
MORPHEME
PSYCH ADJEGI'IVE
kas-
kasmt1
tak-
-Inti+ (ll/ slz)
-Inti+ (II/ siz)
-Inti+ (II/ siz)
iirkliz-
-Inti+ (II/ slz)
-Inti+ (III siz)
dii~iln-
Table 5.
dil~iinttilil
takmtth
takmtJSIZ
iirkiinttilii
iiziin tli 1ii
Psych AdJeCtiVes Denved Wlth --Inti+ (W slz)
-·~
..·
187
-Inr; + l/1 slz also derives psych adjectives which are the modifiers of the
Experiencer argument except for the last two which can also be used as the modifiers of
the Stimulus:
INPUT VERB
MO~~,.
bil-
-In~
+ lii slz
bilin~li
bilin~siz
inan-
-In~
+ 11/ slz
inan~h
inan~stz
iste-
-In~
+ lif slz
PSYCH ADJECIDIVE .
.
'
'
,__,_, __ ,-
·..."i>
isten~li
isten~siz
-In~+
korksev-
-In~+
Table 6.
-/~
korkun~
lif slz
111 slz
Psych Adjectives Denved WJth
sevin~li
-In~+
llJ slz
+ l/1 slz and -It + l/1 slz are other morphemes which derive Experiencer
modifying psych adjectives:
INPUT VERB
anla-
MORPHEME
-I~+ 11/ slz
kavra-
-1~
+ 11/ slz
var-
-1~
+ 11/ slz
PSYCH ADJECTIVE
anlayt~h
anlayt~stz
kavrayt~h
kavrayt~stz
Table 7.
INPUT VERB
urn
Psych Adjectives Denved w1th
MORPHEME
-It+ IV slz
Table 8.
van~h
-I~+
"clever"
ll/ slz
PSYCH ADJECTIVE
urnutlu
urnutsuz
Psych Adjectives Denved WJth -It+ IV slz
-Ar + l/1 slz, -cA + l/1 s/z and -DA + li also derive the following psych
adjectives which modify the Experiencer:
INPUT VERB
duyurn-
-Ar + IV slz
-Ar +II/ slz
Table 9.
INPUT VERB
dU~i.in-
duyarh
duyarstz
umarstz
Psych AdjeCtives Derived with -Ar + IU slz
MORPHEME .
-cA +IV slz
PSYCH A.p}E<:;Jl\'E
di.i~Unceli
dU~Uncesiz
Table !0. Psych Adjectives Derived with -cA +III slz
.···
188
JNPUTV~R.Jl
sev-
sevdah
There is only one psych adjective in our data which seem to be derived from a
psych verb with the morpheme -sf:
INPUTVE_tm
sin-
• MORPIQ!JME
-sl
sinsi
Table 12. Psych Adjectives Derived with -sl
Similarly, there is just one example for the following two psych adjectives both
of which modify the Experiencer:
INPUT VERB
MORPHEME
PSYCH AD.J ECTIVE
yan-
-Ar doner
yanardoner
Table 13. Psych Adjectives Derived with-Ar dOner
INPUT VERB
MORPHEME
PSYCH ADJECTIVE
~a~kaloz
Table 14. Psych Adjectives Derived with -kaloz
-kaloz
The above analysis shows that there are fourteen morphemes which derive
psych adjectives modifying the Experiencer from verbal roots.
On the other hand, the following morphemes derive psych adjectives which
modify the Stimulus argument which as we mentioned can be one of the following:
a. a simple NP referring either to an animate or an inanimate thing
b. an act
c. an event
-DIRI -t + lei and Ill n + mA + d!k are the only morphemes which derive just
Stimulus modifying psych adjectives from psych verbs:
INPUT VERB
MORPHEME
PSYCH ADJECTIVE
bezal9alaldanbtkbunaldaral-
-DIR/ -t + lei
-DIR/ -t+ lei
-DIR/ -t + lei
-DIR/ -t +lei
-DIR/ -t +lei
-DIR/ ct +lei
bezdirici
al9alttct
aldattct
btktmet
bunaltiei
daralttet
189
-DIR/ -t +lei
-DIR/ -t + lei
-DIR/ -t + lei
-DIRI -t +lei
-DIR/ -t + lei
-DIR/ -t + lei
-DIR/ -t + lei
-DIR/ -t + lei
-DIR/ -t + lei
eglengtieengtiveninankankt~ktr-
tirperus anyam I-
eglendirici
giieendiriei
giivendiriei
inandmet
kandmet
kt~klrttel
iirpertici
usandmct
yamltlet
Table 15. Psych AdjeCtives Denved With -DIRI -t +lei
INPUT VERB
MORPHEME
PSYCH M)JECTIVE
bil-
- llJ n + rnA + dlk
tamurn-
- llJ n + rnA + dlk
- llJ n + rnA + dik
bilindik
bilinrnedik
tanmdtk
umulduk
umulmadtk
Table 16. Psych Adjectives Denved With -II/ n +rnA+ dlk
In the third group, there are some morphemes which allow the derivation of
both the Experiencer and the Stimulus modifying adjectives such as -An and -// + rnA +
m/~:
INPUT VERB
MORPHEME
PSYCH ADJECTIVE
uyarsev-
-An
-An
uyaran "stimulating"
seven
Table 17. Psych Adjectives Denved w1th -An
INPUT VERB
MORPHEME
all~-
-II+ rnA+
sev-
-II+ rnA+ ml~
ml~
PSYCH ADJECTIVE
a!J~!lmt~
all~tlmamt~
sevilmi~
sevilmemi~
Table 18. Psych Adjectives Denved With -II+ rnA+ ml~
While the second adjective derived with -DIRI -t + Ik + CI modifies the
Experiencer, the first one with -DIRI -t + Ik modifies the Stimulus:
uy-
tiyduruk
uydurukyu
Table 19. Psych Adjectives Derived with -DIRI -t + lk + Cl
Similarly, the following first adjective refers to a Stimulus while the others
generally refer to Experiencers:
190
The last adjective that follows, modifies usually an Experiencer but the rest
modify the Stimulus:
Except for the first and the last adjectives derived with -Iml nl GI + sAl, all of
the others modify the Stimulus:
Apart from fourteen morphemes which derive psych adjectives modifying the
Experiencer from verbal roots, we identified .two morphemes which derive psych
adjectives referring just to Stimulus. The analysis has also shown that there are six
morphemes which allow the derivation of the adjectives modifying both the Experiencers
and the Stimulus. The next section analyzes the psych adjectives derived from adjectival
roots.
191
II. 5. 2.
Psych Adjectives Derived from Adjectival Roots
The above classification dividing the adjectives as modifying either
Experiencers or Stimulus is also valid for the adjectives which have adjectival roots (which
are very few in number).
In the first group there are Experiencer modifying adjectives deriyyd
wi~h
-
msAr, -/Ik and -C/:
In the second group, there are psych adjectives which modify only the
Stimulus:
II. 5.
3~
Psych Adjectives Derived from Nominal Roots
A similar classification shows that adjectives in this class correspond to the
mentioned three groups. All of the following adjectives are derived from psych nominals.
In the first group, there are four morphemes which derive psych adjectives
which modify the Experiencer. -CL -In and -kar are the most productive ones:
192
The less productive one 1s the following, with which we have only one
example:
The second group of morphemes produces Stimulus modifying psych adjectives
such as -i, -sAl and -DAn:
193
In the last group, there are morphemes which derive both Stimulus and
Experiencer modifying psych adjectives:
In this group, there are two morphemes which produce a great number of psych
adjectives of this kind. These are the most productive psych adjective deriving morphemes
from nominal roots; -ll and -slz. These morphemes are the positive- negative counterparts
of each other and almost all nominals which allow one of them allow the other too to be
attached:
194
-11
-11
-11
-11
-11
-11
sinir
~liphe
tutku
limit
us
zevk
sinirli
~Upheli
tutkulu
limitli
uslu
zevkli
Experiencer
Both
Experiencer
·Both
Experiencer
Both
Table 37. Psych Adjectives Denved wtth -11
II. 5. 4.
Non- Derived Psych Adjectives
In this group, there are psych adjectives which are non derived with any of the
mentioned morphemes. All of these verbs except garip modify only the Experiencer:
195
ters
ukala
uyuz
vakur
yaltak
zevzek
+
+
+
+
+
+
Table 39. Non- denved Psych Adjectives
Mostly these verbs have verbal counterparts with either
these a~zk can be verbalized with either of the helping verbs
-/A~
or -IAn. Among
et or ol and sersem is
verbalized with -/A. Kiis has the same form with its verbal counterpart and vakur can not
be verbalized with any of the helping verbs.
II. 5. 5.
Compound Psych Adjectives
Adjectival compounds have also lexical counterparts. Some of them are derived
with-IA.r:
Some of these verbs are derived with -/An:
196
Some are derived with -!A:
A few number of complex nucleus verbs are first adjectivalized and then
verbalized with -IA.J:
The following verbs are lexicalized with none of the mentioned morphemes and
a.Jzk ol- does not have a lexical counterpart:
197
CONCLUSION
The study aimed at providing a descriptive account of some structural and
semantic aspects of Turkish psych verbs which were not thoroughly analyzed for their own
sake in Turkish.
Section I. 1. summarized the basic discussions about the argument structure
focusing first on the transitivity and the unaccusative phenomenon and second, on the
voice alternations and the causativity which are the basic subjects discussed in the psych
verb literature.
Section I. 2., presented a brief summary of the approaches to psych verbs under
two general headings; structural approaches and semantic/ aspectual approaches.
Section II. 1. gave a brief introduction to psych verbs in Turkish. An analysis of
the Dictionary of Turkish Language Institute (1988) partially provided the data needed to
create a database of psych verbs in Turkish which answers the first research question of the
study.
In this section, first, the criteria of being a psych verb, the properties of the
'
'
Experiencer and the Theme arguments of a psych verb were identified. Then, the inputs of
psych verb verbalization which are non- derived bases, derived verbs and compound forms
were exemplified with the verbs from the data. Lastly, the secondary psych uses in Turkish
which have both psych and non psych senses were also exemplified.
198
Section II. 2. proposed a classification for psych verbs in .Turkish. According to
this classification; the thematic roles, the case marking and the syntactic position of the
arguments have shown that there are four types of psych verbs in Turkish. In this section
these types were introduced and exemplified with an analysis of their inputs of
verbalization. Thus the second hypothesis of the study which claims that Turkish psych
verb classes have similar properties with the ones proposed in the literature for different
languages was proven to be true.
Section II. 3. analyzed the interaction of voice markers and psych verbs in
Turkish according to their transitivity, causativity, passivization, reflexivity and
reciprocity. The third hypothesis of the study that the exceptional behavior of psych verbs
are also observed in their interaction with voice markers in Turkish parallel to the universal
tendencies was also proven to be true.
Transitivity analysis shows that only the Type 1 and Type 2 psych verbs are
transitive and the others are intransitive.
Section II. 3. 2. focused on the interaction between the causativity and psych
verbs in two main sections; morphological causativity and' periphrastic causativity. It was
exemplified that psych verbs similar to other verb classes have their own five sub classes
with respect to morphological causativity. Some psych verbs can morphologically be
causativized, some can not; some psych verbs have psych meanings only in their causative
form, some have only in their non causative form; and lastly, some psych verbs have
different meanings in their causative and non causative forms which are both psych.
199
Causativity analysis shows that Type 2 verbs have an inherent causative
meaning. However the restriction on the further causativization of these verbs is neither
phonological nor semantic but rather morphological. Other than type 2 verbs themselves,
some of the Type 1, most of the Type 3a and all of the Type 3b Type 4 verbs behave just
like Type 2 verbs in their causative forms. These verbs can either be regarded as instances
of different causative variants of their own types or as separate lexical entries which belong
to Type 2.
Periphrastic causativity shows that these verbs have different semantic
interpretations from the causative interpretations of the true members of Type 1, Type 3a
and Type 3b. Sag!a- constructions mostly refer to positive psych events on the part of the
Experiencer and neden ol- constructions refer to rather negative ones. Moreover, sag/aconstructions with the implication of a conscious effort are more compatible with mental
events while neden ol- constructions which imply a non conscious/ accidental cause are
more compatible with emotional ones ..
This section has also analyzed the semantic notions of control and the
permissive reading and concluded that if a psych event is not self/ internally controll!lble, it
can not be controlled by an outer/ external agent and vice versa. If the psych event is a
controllable event, it can be causativized but if the event is a non-controllable one, then the
causative form seems rather odd. If the event is a non controllable one but there is still a
causative form, this implies that there is a physical act or a series of physical acts involving
in the causative event to evoke the emotional state on the Experiencer. Moreover, it was
concluded that when the morphological counterpart of a psych verbs is negated, the
periphrastic paraphrase has a permissive reading.
200
Sections II. 3. 3. and II. 3. 4. show that most of the psych verbs with the passive
morpheme -Il have a reflexive meaning and the argument indicated with by phrase is also
a kind of Causer of the event. However, this argument does not actively participate in the
event since it is just a Stimulus for the Experiencer.
Section II. 3. 5. has shown that although a few number of psych verbs in the
data allow the attachment of the reciprocal morpheme, in general, these verbs do not have
a reciprocal meaning. Semantically, this is due to the reason that psych activities are
personal matters which can not be done reciprocally or cooperatively.
Section II. 4. shows that in Turkish, psychological states are expressed with
compound forms as well as by lexical means. There are a number of compound psych
verbs in Turkish which display similar structural and semantic properties with their nonps"ych counterparts which was the fourth hypothesis of the study proven to be true.
Turkish uses 31 helping verbs for this purpose. Most of the time, the lexical
form and its compound counterpart have the same meanings. According to their
components, compound psych verbs can be derived with a nominative, accusative or a
dative nominal or they can be derived with a complex nucleus or an adjective.
Due to the argument structure properties of the helping verb, compound forms
may be claimed to add six new classes to our existing classification of psych verbs. The
argument structures of all of the compound forms are consistent with the argument
structure of their non psych counterparts in that they either incorporate the Theme or the
Goal arguments into the compound.
The last section analyzed the derivational properties of the adjectives which
have psychological senses. Morphologically, their input can be a verbal, an adjectival or a
nominal root. Alternatively, they may be non- derived in any way or they may be
201
compound forms. The derivation of psych adjectives does not display any idiosyncratic
properties which differentiate them from general derivational patterns of Turkish as the
fifth hypothesis ofthe study claims.
On the other hand, semantically, there are two main types of psych adjectives to
which a third one which have the both functions could be added; those which modify the
Experiencer argument, those which modify the Stimulus argument and those which modify
both the Experiencer and the Stimulus arguments.
Possible further studies analyzing the application of the proposed syntactic tests
to Turkish psych verbs and also the aspectual analysis of the properties of these verbs are
going to help to gain new insights as to the nature of Turkish psych verbs which still need
to be checked against the crosslinguistic findings.
202
BIBLIOGRAPHY
Alberti, Gabor. 1997. Argument selection. Frankfurt: Peter Lang.
Alsina, A. 1996. The role of argument structure in grammar. California: CSLI.
Arad, Maya. Psych-notes. Access date 10.06 2002, www address:
http ://wwwphon. ucl. ac. uk/home/PUB/WPL/98papers/arad. pdf
Baker, Chris and Dowty David. Nonverbal thematic proto-roles. Access date 18.02.03,
www address:
http ://ling. ucsd.edu/~barker/Researchlbarker.dowty .nominal. proto-roles. pdf
Baker, Chris. On the structural positions of themes and goals. Access date 17.12.2002
www address: http://ling.rutgers.edu/people/facultylbaker/dat-shift-ui-prt.pdf
Belleti, Adriana and Luigi Rizzi. 1988. Psych verbs and theta theory. Natural Language
and Linguistics Theory 6, 291-352.
Bennis, Hans. Unergative adjectives and psych verbs. Access date 23.03.2003, www
address: http :1/wwwmeertens.knawnl/medewerkers/hans.bennis/unacc. pdf
Borik, 0. Experiencers and goals in German unaccusatives. Access date 19.07.2003,
www address:
wwwlet.uu.nl/~Patrick.Brandt/personal/goal.ps
Bresnan, Joan. 1995 Lexicality and argument structure. Access date 11.01.2003, www
~
address: http:// www-lfg.stanford.edu/lfg/bresnan/paris.ps
Chung, Taegoo. Argument structure andEnglish grammar. Access date 02.01.2003,
www address:
http://pc171115.pc.waseda.ac.jp/ccdl/cl korea/0612 handoutl.pdf
Cook, Vivian and Newson, Mark. 1997. Chomsky's universal grammar: An
introduction. Oxford: Blackwell.
203
Cruse, D. A. 1983. Lexical semantics. Cambridge: Cambridge University Press.
Culicover, Peter. W. 1997. Principles pnd parameters: An introduction to syntactic theory
New York: Oxford University Press.
Dabrowska, Ewa. 1994. Dative and nominative experiencers: Two folk theories of mind.
Linguistics 32, 1029-1054.
Dikken, Marcel. 1996. Review of "Verb Incorporation and Elementary Predicates" by
Murat Kural. Glot International2/6, 10-13.
Dowty, David. 1991. Thematic proto-roles and argument selection. Language 67/3, 547619.
Erguvanh Taylan, Eser. An odd case in the causative constructions of Turkish. CLS 16, 92100.
Filip, Hana. 1996. Psychological predicates and the syntax semantics interface. In
Goldberg, Adele (Ed.), Conceptual structure, discourse and language.
Stanford: CSLI.
Frawley, William. 1992. Linguistic semantics. New Jersey: Lawrance Erlbaum.
Georgopoulos, Carol. 1991. On psych predicates. In Georgopoulos, Carol and R. Ishihara
(Eds.), Interdisciplinary approaches to language (217- 238). Dordrecht:
Kluwer Academic.
Goksel, Ash. 1993. Levels of representation and argument structure in Turkish.
Unpublished Ph. D. Dissertation. London University, Lol,ldOn. _
Grimshaw, Jane. 1990. Argument structure. Cambridge: MIT.
Hale, Ken and Keyser, Samuel Jay. 2002. Prolegomenon to a theory of argument structure.
Cambridge: MIT.
Hatori, Yuriko. 1997. On the lexical conceptual structure of psych verbs. In Kageyama,
204
Taro. (Ed.) Verb semantics and syntactic structure (15-44). Tokyo: Kurosio.
Hoekstra, Teun. 2000. The nature of verbs and Burzio's generalization. Reuland, Eric.
(Ed.), Arguments and case: Explaining Burzio 's generalization (57- 78).
Philadelphia: John Benjamins.
Holmes, Jasperr. The syntax and semantics of causative verbs. Access date 22.08.2002,
www ·address: http://wwwphon. ucl.ac. uklhome/PDB/WPL
Kordoni, Evagelia. Argument structure and linking in HPSG: Evidence from Experiencer
psych verb constructions. Access date 03.09.2003, www address:
http://wwwiccs. informatics. ed.ac. uk/ archive/ esslli97 /papers/kordoni/kordoni. ps
Kordoni, Valia. 2001. Linking Experiencer-subject verb constructions in Greek. Access
date 01.07.2002, www address:
http://csli-publications.stanford.edu/HPSG/1/hpsgOOkordoni.pdf
Kordoni, Valia. 2001. Optimal linking for modern Greek psych verb constructions. Access
date 27.10.2002, www address:
http:/I csli-pub lications.stanford. edu/LFG/6/lfgO 1kordoni. PS!_f
Krifka, Manfred. 2001. Specific verb classes and alternations. Access date 25.10.2002,
www address: http ://amor.rz.hu-berlin.de/~h2816i3x/LexSemantik7 .pdf
Kural, Murat. 1996. Verb incorporation and elementary predicates. Unpublished Ph. D.
Dissertation. University of California, Los Angeles.
Levin, Beth 1993. English verb classes and alternations: A preliminary investigation.
London: The University of Chicago Press.
Levin, Beth and Hovav, Rappaport. 1996. Unaccusativity at the syntax semantics interface.
Cambridge: MIT.
Levin, Beth and Hovav, Rappaport. 1996. Lexical semantics and syntactic structure. In
205
Lappin, Shalom (Ed.). The handbook of contemporary semantic theory. Oxford:
Blackwell.
Levin, Beth. Objecthood: An event structure perspective. Access date 06.06.2002, www
address: http://wwwlot.let. uu.nl/zs200 1/papersMarantz/cls99. pdf
Levin, Beth and Hovav, Rappaport. Change of state verbs: Implications for the theory of
argument structure. Access date 24.05.2003, www address:
http://wwwstanford.edu/~bclevinlbls02hdr.pdf
Levin, Beth and Hovav, Rappaport. 2002. The semantic determinants of argument
expression: A view from the English resultative construction. Access date
16.09.2002, www address: www stanford.edul~bclevin/paris02.pdf
Manning, Christopher. D. 1996. Ergativity: Argument structure and grammatical relations.
California: CSLI.
Martin, Fabienne. Object experiencer psych verbs and causatives. Access date
04.10.2003. www address:
http://www -ui lots.let. uu.nl/ conferences/Perspectives on Aspect/Proceedings/
/martin. pdf
Masayuki, Wakayama. The peculiarity ofpsych and impersonal verbs in the loss ofV
moveJllent. Access date 29.11.2003, www address:
http://wwwj uen. ac. jp/english! elmug/study/pdf/waka(2000b). pdf
Me Ginnis, Martha. Semantic and morphological restrictions in Experiencer predicates.
Access date 05.11.2002, www address:
http://wwwucalgary.ca/-mcginnis/papers/CLAOO.pdf
Me Ginnis, Martha. Event heads and the distribution of psych roots. Access date
26.10.2003, www address:
206
http ://wwwucalgary.ca/~mcginnis/papers/PWPL6 3 .pdf
Motomura, Mitsue. Linking problems and predicates of psychological state in Japanese.
Access date 19.11.2003, www address:
http://www ling. umd.edu/Events/StudentConference/2002conference/abstracts/
Motomura.pdf
Mulder, Rene. 1992. The aspectual nature of syntactic complementation. Leiden: Holland
Institute of Generative Linguistics.
Nakamura, Wataru. Psych verb constructions and a double tier theory of argument
structure. Access date 17.11.2003, www address:
www3.aa.tufs.ac.jp/~ritsuko/project/naka981 O.doc
Nakipoglu, Mine. 1998. Split intransitivity and the syntax semantics interface.
Unpublished Ph. D. Dissertation. University of Minnesota, Minnesota.
Nelson, Diane. Linking causatives and Experiencers. Access date 14.10.2002, www
address: http://wwwleeds.ac.uk/linguistics/research
Nilsson, Brigitte. 1985. Case marking semantics in Turkish. Unpublished Ph. D.
Dissertation. University of Stockholm, Stockholm.
Palmer, F. R. 1994. Grammatical roles and relations. Cambridge: Cambridge University
Press.
Payne, Doris L. Tre explanation of argument structure: Lexicalization or discourse use?
Access date 14.08.2003, www address:
http://wwwuoregon.edu/~dlpayne/Explanation.pdf
Pinon, Christopher H. A finer look at causative inchoative alternation. Access date
09.01.03, www address:
http ://wwwphil-fak. uni -duesseldorf. de/~pinon/papers/pinon flcia. pdf
207
Pylkkanen, Liina. 2000. On stativity and causation. In Tenny, Carol and Pustejovsky,
James. (Eds.), Events as grammatical objects: The converging perspectives of
lexical semantics and syntax (417-442). California: CSLI.
Pesetsky, David. 1987. Binding problems with Experiencer verbs. Linguistic Inquiry 18,
126-140.
Pesetsky, David. 1995. Zero syntax: Experiencers and cascades. Cambridge: MIT.
Pinker, Steven. 1995. Affectedness and direct objects: The role of lexical 'serharitics in tfie
acquisition of verb argument structure. In Levin, Beth and Pinker, Steven
(Eds), Lexical and conceptual semantics (153-196). Oxford: Blackwell.
Ravin, Yael. 1990. Lexical semantics without thematic roles. Oxford: Clarendon.
Reese, Brian. J. Improper subjects in Hindi experiencer constructions. Access date
01.11.2003, www address:
http :1/ling. uni-konstanz.de/pages/conferences/salaO 1I abstracts/reese. pdf
Roca, I. M. 1992. Thematic structure: Its role in grammar~ In Hale, Ken and Keyser, J.
The syntactic character of thematic structure. Berlin: Foris.
Sag, Ivan A. and Szbolcsi, Anna (Eds.). 1992. Lexical matters. California: CSLI.
Tenny, Carol. L. 1994. Aspectual roles and the syntax semantics interface. Dordrecht:
Kluwer Academic.
TDK Tiirkye Sozliik. 1998. Ankara: TDK Yaymlar1.
TDK ingilizce- Tiirkye Sozliik. 1999. Ankara: TDK Yaymlar1.
Tietze, Andreas. 1989. Observations on the convergence of passive and medioreflexive
verb forms: The case of modem Turkish. In Wexler, P. et al (Eds.), Studia
Linguistica et Orientalia Memoria Haim Blanc Dedicata (239-252).
Weisbaden: Otto Harrossowitz.
208
Van Valin, Robert, D. Cross linguistic patterns oflinking. Access date 12.10.2002,
www address:
http://linguistics. buffalo. edulresearchlrrg/vanvalin papers/XlingP attLnkg. pdf
Verspoor, Marjolijn. H. 1990. Semantic criteria in English complement selection.
Unpublished Ph. D. Dissertation. Leiden University, Leiden.
Vanhoe, H. 2002. Aspects ofthe syntax of psychological verbs in Spanish: A lexical
functional analysis. (In) Butt, M. and King, H. T. (Ed.), Proceedings of the
LFG 02 Conference (373-388). California: CSLI.
Voorst, Van. J. 1992. Aspectual semantics of psychological verbs. Linguistics of
Philosophy 15, 65-92.
Wechsler, Stephen. 1995. The semantic basis of argument structure. California: CSLI.
Wierzbicka, Anna and Harkins, Jean. 2001. Emotions in crosslinguistic perspective.
Berlin: Mouton de Gruyter.
Williams, Edwin. 1981. Argument structure and morphology. The Linguistic Review 1, 81114.
Wunderlich, Dieter. Argument linking types for transitive verbs. Access date 19.07.2003,
www address: http://web.phil-fak.uni-duesseldorf.de/-wdl/arg-link-types.pdf
Zimmer, Karl. 1988. Semantic motivation in case assignment: Another look a:t Turkish
causatives. (In) Ko<;, Sabri (Ed.), Proceedings of the Fourth International
Conference on Turkish Linguistics (217-223). Ankara: METU.
Pinar İbe Akcan