PROTOTYPE THEORY AND

THE MEANING OF VERBS,

WITH SPECIAL REFERENCE TO

MODERN GREEK VERBS OF MOTION

by

Eleni Antonopoulou

A t h e s i s submitted f o r the degree o f

Doctor o f Philosophy

School o f O r i e n t a l and A f r i c a n Studies

U n i v e r s i t y o f London

1987
 ProQuest N um ber: 10731450

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is d e p e n d e n t upon the quality of the copy subm itted.

In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript

and there are missing pages, these will be noted. Also, if m aterial had to be rem oved,
a n o te will ind ica te the deletion.

uest
ProQuest 10731450

Published by ProQuest LLC(2017). C opyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States C o d e
M icroform Edition © ProQuest LLC.

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 4 8 1 0 6 - 1346
 ABSTRACT

The present study tests the ap plicability of Prototype t he o ry ,

sel ect ed among competing theoretical frameworks, to a lexical
semantic anal ys is of verbs, wi th particular reference to the
p r ev i o u s l y uncharted domain o f Modern Greek verbs of motion. A
number o f the c h a r a c t e r i s t i c s which Prototype theor y e st a b l i s h ed in
connection wi th c e r t a i n types of nouns are demonstrated to p e r t a i n to
verbs: their meaning is not a m a tt e r of necessary and sufficient
c o n d i t i on s, but rather a matter of g r a da t i o n; their attributes
combine in non-arbitrary ways to form c at e g o r i e s with f uzzy
boundaries the members of which are non-equivalent. Two
categorizations o f motion verbs according to ' ma jor classificatory
properties' are discussed at length. First, 'sta te s', 'processes'
and ' e v e n t s ' are shown t o c o n s t i t u t e a continuum, t he focal points o f
which are identifiable on the basis o f the interaction of factors
such as s pat io-te mpor al specifications, aspect, i n h e re nt semantic
properties of individual verbs and the nature o f the 'theme' (moving
object). Second, 'causativity' and 'agentivity' are understood as
distinct, to some e x t e n t , c l u s t e r s o f s c a l a r p r o p e r t i e s and d i f f e r e n t
Modern Greek motion verbs are shown to e x h i b i t these p r o p e r t i e s to a
g r e a t e r or l e s s e r degree. In seeking to determine which f a c t o r s may
be r esponsi bl e f o r the formati on of verb c a t e g o r i e s , it is recalled
th at the v a l i d i t y o f the p r i n c i p l e of 'family resemblance' and the
method for identifying the 'basic' level .of abstraction cannot be
te s te d in the case o f verbs. It is suggested t h a t o t he r f a c t o r s may
be o p e r a t i v e , such as the r e l a t i v e ' s a l i e n c e ' o f c e r t a i n combinations
of properties, 'lin g u is tic markedness', fam iliarity and frequency.
This t e n t a t i v e conclusion i s r e i n f o r c e d wi th r espect to Modern Greek
verbs o f motion by the r e s u l t s o f s p e c i f i c t e s t s .

- 2 -
 ACKNOWLEDGMENTS

I wish to express my g r a t i t u d e to my s uper vi sor Dr. D.C. Bennett f o r

hi s invaluable guidance and advice and hi s in fin ite p at ien ce and
constant encouragement. I am also deeply indebted to Professor D.
Theophanopoulou-Kontou for her i ne st ima bl e intellectual stimulation
and personal kindness.

My thanks are also due to several friends and c o ll ea gues , too

numerous to mention here, f o r advice and much needed moral support.
P a r t i c u l a r mention must be made o f Miss A. K ok k ol i, who helped wi th
several pa rt s o f t h i s manuscript, o f Dr. S. Asproudis, who a ssi st ed
in the sta tistic a l a na ly s is of test results, and of Miss A.
Tsomlektsoglou, who demonstrated g r e a t s k i l l arid d i l i g e n c e in p u t t i n g
t h i s t h e s is i n t o i t s f i n a l form.

^ N o t e : T h e m a j o r w o r k for t h i s d i s s e r t a t i o n w a s c a r r i e d
out d u r i n g my y e a r s at S O A S a n d w a s s u b s t a n t i a l l y c o m p l e t ­
ed b e f o r e 1983.

- 3 -
 CONTENTS

Page

INTRODUCTION........................................................................................................................ 10

1. THEORETICAL FRAMEWORKS FOR A LEXICAL SEMANTIC ANALYSIS

OF MOTION VERBS...........................................................................................................13
1.1 A case f o r l i n g u i s t i c l e x i c a l semantics...........................................13
1 .2 Existing the o re tica l frameworks: a d i s c u s s i o n ............................. 20
1 . 2 . 1 Componential a n a l y s i s : the ' m a j or ' shortcomings............ 20
1.2.1.1 Nature of f e a t u r e s .................................... . . . . 2 0
1.2.1.2 Markedness........................................................................24
1.2.1.3 A to m i c i t y and u n i v e r s a l i t y o f f e a t u r e s 25
1 . 2 . 2 Componential a n al y s i s and semantic f i e l d t h e o r i e s
viewed as s t r u c t u r a l i s t t h e o r i e s .......................... 28
1 . 2 . 3 The e mp ir i ca l v a l i d i t y o f componential a n alysis., 29
1 . 2 . 4 Componential a n al y s i s and semantic f i e l d t h e o r i e s :
the ' mi n o r ' shortcomings............................................... 31
1 . 2 . 5 S t r u c t u r a l i s m , d e s c r i p t i v e adequacy, and the case
o f motion v e r b s .............................................................. 33
1 . 2 . 6 S t r u c t u r a l i s m vs Prototype theory:
c o n c e p t u a l i z a t i o n and e x t r a ! i n g u i s t i c r e a l i t y ...,4 5
1.3 Prototype t he or y and human c a t e g o r i z a t i o n ...................................... 49
1 . 3 . 1 On Putnam's s t e r e o t y p e s .................................................................. 49
1 . 3 . 2 On Rosch's p r o t o t y p e s .......................................................................59
1.3.2.1 Stereotypes and prototypes: common
ground.............................................. 59
1.3.2.2 Prototype theory as a theory o f
c a t e g o r i z a t i o n . ............................................................61
1.3.2.3 The basic l e v e l of a b s t r a c t i o n ........................... 62
1.3.2.4 Category f o r m a t i o n . . . . . ......................................... 64
1 .4 D e l i m i t a t i o n o f the f i e l d o f motion v e r b s .......................................70
Notes on Chapter 1 .................... ...7 8

. 4 -
2. THE 'STATES-PROCESSES-EVENTS' CATEGORIZATION AND ITS
APPLICATION TO MOTION VERBS................................................................................ 82
2.1 P r e l i m i n a r i e s .......................................................................................................82
2 .2 An overview o f the standard t e s t s f o r d i s t i n g u i s h i n g
between s t at e s- p ro c e s s e s - e v e n t s and howt o f a i l them................. 83
2 .3 Fur the r c r i t e r i a f o r the d e f i n i t i o n o f
s t a t e s - p r o c e s s e s - e v e n t s , ..............................................................................88
2 .4 Adv er bi al s o f time and goal vs l o c a t i v e ............................... 93
2 .5 Aspect, s t a t e s - p r o c e s s e s - e v e n t s , and Modern Greek
motion v e r b s .........................................................................................................97
2 .6 Nature o f the t h e m e . . .................................................................................. 102
2 .7 Concluding remarks on the r e d e f i n i t i o n o f
s t a t e s - p r o c e s s e s - e v e n t s .............................................................................. 104
2 . 8 Test f rames...................... 108
2 .9 Comments on L i s t I I I . . ............................................................ 110
Notes on Chapter 2 ............................................................................. 116

3 . THE 'CAUSATIVITY-AGENTIVITY' CATEGORIZATION AND ITS

APPLICATION TO MOTION VERBS..............................................................................117
3.1 How separate can the notions o f ' c a u s a t i v i t y ' and
'agentivity' be k e p t ? .................................................................................. 117
3 . 2 I d e n t i f y i n g Modern Greek causati ves o f moti on............................ 127
3.2,1 Non-causative t r a n s i t i v e s .........................................................,138
3 . 3 A g e n t i v i t y as a c l u s t e r o f p r o p e r t i e s .............................................. 143
3 . 4 T r a n s i t i v i t y and a g e n t i v i t y .....................................................................150
3 . 5 D i r e c t vs i n d i r e c t c a u s a t i v e s ................................................................ 157
3 . 6 A g e n t i v i t y measurements f o r i n t r a n s i t i v e s ..................................... 170
3 . 7 Modern Greek mediopassives and passives as
i n t r a n s i t i v e s ........................ 176
Notes on Chapter 3 ................................................................................ 189

- 5 -
4. PRINCIPLES OF CATEGORIZATION AND MINOR PROPERTIES OF MOTION
VERBS............................................................................................................................... 194
4.1 P r i n c i p l e s o f c a t e g o r i z a t i o n o f motion v e r b s . .............................194
4 . 1 . 1 Taxonomies f o r v e r b s . .....................................................................199
4 . 1 . 2 Levels o f i nc lusi veness and l i n g u i s t i c a l l y
unmarked c a t e g o r i e s ......................................................................... 204
4 . 1 . 3 Taxonomic sets proposed f o r Modern Greek
motion v e r bs ................................................................................ ,209
4 . 2 Minor p r o p e r t i e s o f Modern Greek motion v e r b s ............................ 224
4.2.1 ' C h a n g e - o f - l o c a t i o n ' and ' d i r e c t i o n a l i t y ' ........................227
4.2.2 ' P a t h ' and 'dependent m o t i o n ' ...................................................231
4.2.3 'Change o f o r i e n t a t i o n ' . . ............................................................234
4.2.4 ' M a n n e r ' , 'medium', and ' i n s t r u m e n t a l i t y ' . . ......................237
4 . 3 Hi er ar chy o f p r o p e r t i e s : taxonomies and paradigms......................250
4 . 4 Motion verbs and the n o n - a r b i t r a r i n e s s o f c a t e g o r i e s .............258
Notes on Chapter 4 ..................................................................................................252

5. ELICITING INFORMATION FROM NATIVE SPEAKERS............................................ 265

5.1 Semantic s i m i l a r i t y s o r t i n g tasks and c l u s t e r a n a l y s i s 265
5 . 1 . 1 Semantic s i m i l a r i t y t e s t s i n v o l v i n g English
motion v e r b s ..................................................... ...................... .......... 268
5 . 1 . 2 A semantic s i m i l a r i t y t e s t wi th 34 Modern Greek
motion v e r b s ........................ 278
5 .2 P r o t o t y p i c a l i t y t e s t s w i t h Modern Greek motion v e r b s .............. 295
Notes on Chapter 5 ...................................... 318

Concl usions.................. 321

Appendix: List I Modern Greek verbs o f motion and

p o s i t i o n ......................................................................... 326
List I I Five t e s t frames f o r Modern Greek verbs
o f mot ion................................... ................................. 332
List I I I A c l a s s i f i c a t i o n o f Modern Greek verbs
o f motion along the process-event
continuum ....................................................... 335

- 6 -
 L i s t IV Relative a g en tivity o f sel ect ed
c au sa ti ve Modern Greek verbs o f m o t i o n . . . 3 3 6
List V Relative a gentivity o f s el ect ed non­
c au sa ti ve Modern Greek verbs o f m o t i o n . . . 341
L i s t VI Taxonomic sets and na tur al classes o f
Modern Greek verbs o f moti on..............................349
List V II A c la s s i f ic a t i o n of Modern Greek verbs
o f c h a n g e - o f - p o s i t i o n ............................................383
List V I I I Modern Greek verbs o f c h a n g e - o f - l o c a t i o n
presented in diagrammatic f or m.......................... 384

BIBLIOGRAPHY...................................................................................................................... 385

- 7 -
 ABBREVIATIONS, NOTATION AND FORMAT OF EXAMPLES

A full l i s t i n g o f the a b b re vi a t i o ns used in t h i s t e x t is given below:

Adv. = a d ve r b/ a d ve r bi a l
Adv.Phr. *■adve rbi al phrase
CA = c l u s t e r a n al y s i s
CAUS = c aus at ive
CC = cross-classification
CL = change o f l o c a t i o n
CP = change o f p o s i t i o n
DC = d i r e c t causation/causative
EC = e x p l i c i t c au sa ti ve
FS = Formal Semantics
HCS = hierarchical s t r u c t u r i n g scheme
IC = in d ir e c t causation/causative
If. « im perfect/imperfective
INTR = intransitive
LC = lexical c au s at iv e
LOC = locative
MG - Modern Greek
MGMV = Modern Greek motion verb
MV = motion verb
NC = n a t ur a l class
NON-CAUS - non-causati ve
NP = noun phrase
N^LOC “ noun phrase o f l o c a t i o n
Pf. = perfect/perfective
PP = prepositional phrase
Prep. = preposition
S = sentence
SM * s el f- movi ng
s .o . * someone
S-P-E = state(s)-process(es)-event(s)
SST = semantic s i m i l a r i t y s o r t i n g t as k
s.th. = something

- 8 -
 s.wh. = somewhere
TR = transitive
V = verb
VP = verb phrase

Phonol ogi ca ll y identical verbs appear as separ ate items wi th

subscr ipts (e.g. 'ro ll}', 'ro llg '). The s ubsc ri pt s are used to
fa c ilita te t he d e s c r i p t i o n and are not meant as an indication that
d i f f e r e n t items are i nvolved r a t h e r than a s i n g l e polysemous one.

Modern Greek examples appear in broad phonemic transcription. The

following conventions are adopted: palatalization of / k/, /v /, /x /
and / g/ before front vowels is not marked; non-syllabic <i> is
t r a n s c r i b e d as / j / ; p r e n a s a l i s a t i o n o f voiced p l o s i v e s i s not marked
systematically. Titles o f Greek newspapers, magazines and books and
a uthors' names are transliterated, as are also t he terms
' katharevousa', 'dim otiki' and 'Koine Nea E llin ik i'. Modern Greek
examples are f ol l ow ed by an English t r a n s l a t i o n ; a word-by-word gloss
i s also provided whenever t h i s i s considered necessary.

9 -
 INTRODUCTION

This t h e s i s attempts a l e x i c a l semantic a nal ys is o f motion verbs, and

in p a r t i c u l a r Modern Greek verbs o f motion, w i t h i n the framework of
Prototype t h e o r y . While t he method o f approach adopted here has been
e x t e n s i v e l y t e s t e d in domains covered by nouns, i t has not p re vi o us l y
been appl ie d s y s t e m a t i c a l l y to verbal categories. As a r e s u l t , the
a na ly s is is both t h e o r e t i c a l and d e s c r i p t i v e , o f n e ce s s i t y touching
upon psycholinguistic theories of human categorization and
p hi l osop hi ca l ' discussions of the relation between linguistic
expressions and e x t r a l i n g u i s t i c r e a l i t y .

Prototype t he or y - which is adopted in this study as a more

a p p r o p r i at e method than componential a na ly s is and semantic field
theories - has raised a number of interesting issues which are
examined here i n connection w it h Modern Greek motion verbs. The main
p r i n c i p l e s i nvolved are the f o l l o w i n g :

- Word meaning is not a ma tt er of necessary and sufficient

conditions.
- Hyponyms are not e q u i d i s t a n t from t h e i r super or di na te ; the members
of a category are not e q u a l l y r e p r e s e n t a t i v e o f t h i s cat egory.
- Semantic properties are not a rb itrarily combined to form
categories.

The most i mportant general issue, in my opi ni on, is g ra da t io n and the

f uz z in es s of boundaries between semantic categories. All these
t e n e t s o f Prototype t he or y seem e q u a l l y r e l e v a n t t o the d e s c r i p t i o n
o f the semantic f a c t s o f verbs as well as o f nouns.

Motion verbs constitute a fa irly well-defined semantic field and

exhibit a number o f interesting properties also relevant to ot he r
verbal domains. The f i e l d has the a d d i t i o n a l m er it s of being hi g hl y
structured, of c o n ta i ni n g many cat egor ie s fam iliar to most native
speakers and, what is more i mportant, of involving c at egor i es more

- 10 -
r e a d i l y de sc r ib a bl e i n terms o f perceptual and f u n c t i o n a l properties
than those o f most o t h e r verbal domains. This i s perhaps the reason
why a l o t o f l i n g u i s t i c research has been conducted in t h i s area of
the vocabulary of languages ot he r than Modern Greek, especially
English and German.

Modern Greek verbs o f motion have never been s y s t e m a t i c a l l y examined,

which has n e c e s s i t a t e d the i n c l u s i o n in t h i s study o f a consi der able
volume of data. The peculiar sociolinguistic situation of Modern
Greek, namely the intermingling of ' katharevousa' (t he 'p u ris tic'
language v a r i e t y ) and ' d i m o t i k i ' ( t he ' c o l l o q u i a l ' / ' p o p u l a r ' variety)
presents special problems in several areas, from the collection of
data and the f or mati on o f taxonomic s t ru c t u r e s to t he assessment of
p r o t o t y p i c a l i t y judgments o f s u b j ec t s.

Some o f the te n et s o f Prototype t he o r y , such as the i d e n t i f i c a t i o n of

the 'basic level of abstraction' and the principle of 'fam ily
resemblance' f o r cat egory formati on cannot be r e a d i l y t e s t e d in the
case of verbs. A verification of these t ene ts presupposes the
possibility of e lic itin g attributes (properties) of cat egor ie s
directly from n a t i v e speakers, which i s not e v id e nt in t he case o f
verbs.

Never the le ss , it can be shown t h a t n a t i v e speakers do make r e l i a b l e

judgments on the r e l a t i v e d i st an ce between the i n c l u s i v e category and
i t s hyponyms i n those cases where such taxonomic o r g a n i z a t i o n is well
established. An i mportant p o i n t is t h e r e f o r e r a i s e d concerning what
such judgments are based on, i.e. what is r esponsi bl e for the
formati on o f c at e g o r i e s in the case o f verbs, and in the area under
investigation in p a r t i c u l a r . This study does not a sp i re t o provide
answers to such i mportant problems, especially as i t is a piece of
linguistic rather than psycho!inguistic lexical semantics. It
cannot, however, avoid testing the validity of certain proposals
concerning the forma ti on o f c a t e g o r i e s , such as semantic s i m i l a r i t y ,
number o f shared p r o p e r t i e s , the r e l a t i v e s a l i e n c e o f combinations o f
properties, and the natur e o f the 'b asic' level of abstraction for
verbs.

- II -
The s t r u c t u r e o f the present study i s as f o l l o w s :

Chapter I discusses at some l ength the search for an a pp r op ri at e

theoretical framework for the description to follow. It also
i ncludes a s ecti on on t he d e l i m i t a t i o n of the f i e l d o f motion verbs.

Chapters 2 and 3 contai n an account of those properties of verbs

which are here considered 'maj or c l a s s i f i c a t o r y properties', namely
s t a t e s - p r o c e s s e s - e v e n t s , c a u s a t i v i t y and a g e n t i v i t y .

Chapter 4 discusses the relations between Modern Greek verbs of

motion and the 'mi nor properties' considered relevant for the
d e s c r i p t i o n o f the verbs in quest ion. It provides an answer to the
question o f how the f i e l d i s organi zed.

Chapter 5 r e p or t s the results of tests conducted wi th a view to

checking the p os si bl e psychological r e a l i t y o f c e r t a i n p r o p e r t i e s and
s t r u c t u r e s based on the 'purely l i n g u i s t i c ' a n al y s i s c a r r i e d out in
the preceding chapt er s. I t also includes an attempt t o i d e n t i f y some
of the factors r espons ibl e for the formation o f protot ypes in the
domain o f Modern Greek verbs o f motion.

An Appendix is i ncluded which contains six lists of Modern Greek

motion verbs. List I comprises 181 e n t r i e s which are indicative of
the m a t e r i a l the description is based on. Besides motion verbs a
number o f verbs o f p o s i t i o n are also i ncluded, speci al r ef erence to
which i s made in Chapters 2 and 4. L i s t I I contains f i v e t e s t - f r a m e s
and the verbs o f L i s t I which can occur in each one o f them. List
III presents a c l a s s i f i c a t i o n o f Modern Greek motion verbs along the
'process-event' continuum. Lists IV and V show the r e l a t i v e degree
o f a g e n t i v i t y o f a number o f caus ati ve and non-causati ve Modern Greek
verbs o f motion. L i s t VI contains a l l the taxonomic sets and natur al
classes identified w i t h i n the f i e l d under i n v e s t i g a t i o n . Lists V I I
and V I I I present a c l a s s i f i c a t i o n o f motion verbs based on the 'minor
p r o p e r t i e s ' they e x h i b i t .

- 12 -
 1. THEORETICAL FRAMEWORKS FOR A LEXICAL
SEMANTIC ANALYSIS OF MOTION VERBS

1.1 A case f o r l i n g u i s t i c l e x i c a l semantics

In an area of scientific investigation where there is little

agreement on what the subj ec t ma tte r r e a l l y is, it i s no wonder t h a t
there is no agreement on its pa rt s and their c o n te n t. * Semantic
t heory lends its elf readily to such confusion by lying at the
crossroads o f ' p u r e ' l i n g u i s t i c s , psychology and l o g i c .

Logical grammar or Formal Semantics (hencef or th FS) understands and

t h e r e f o r e descri bes the semantics o f natur al language in terms of a
theor y of entailments. Consequently it sets its elf goals such as
matching meanings to syntactic categories, d e sc r ib i ng sentences
through their truth-conditions or e xp l a i n i n g how t he meanings of
individual words contribute t o the o v e r a l l meaning o f the sentence
they belong t o . So word meaning i s not t o be analysed in its own
terms but only in terms o f t h i s c o n t r i b u t i o n . This i s a l l very well
f o r ' l o g i c a l ' words (such as ' a n d ' , 'but', 'o r', 'necessarily'). The
v as t m ajority of non-logical words is, however, le ft out, in the
sense t h a t FS refuses t o break i n t o t h e i r semantic c o n te n t. Lexical
meaning is relegated to l exi cogr aphy . This is very simply and
e x p l i c i t l y expressed i n Thomason ( 1 97 4 : 4 8 ) :

"the problems o f semantic t heory should be d i s t i n g u i s h e d from

those o f l e x i c o g r a p h y . . . we should not expect a semantic theory
to f u r n i s h an account o f how two expressions belonging to the
same s y n t a c t i c category d i f f e r in meaning. 'Wal k' and ' r u n ' ,
f o r i n s t a n c e , . . , c e r t a i n l y do d i f f e r in meaning, and we r e q ui r e
a d i c t i o n a r y o f English t o t e l l us how. But the making of a
d i c t i o n a r y demands c onsi der able knowledge o f the wor l d".

Cresswell ( 1 9 78 : 4) considers the same example o f two d i f f e r e n t but

related words, namely 'walk' and 'ru n ', only to explain that the
s peaker' s knowledge t h a t the two verbs share a +movement component

*A notable exception is D o w t y ( 197 9).

- 13 -
i s not t o be i n cor por at ed i n the m o d e l - t h e o r e t i c apparatus because i t
i s "a lre ady present in the func ti ons d e sc r ibi ng ' w a l k ' and ' r u n ' " .

In simple terms, there is such a t h i ng as purely linguistic

knowledge, which i s compl etel y d i s t i n c t from knowledge o f the world
and the l a t t e r is the concern o f l e x i c og r ap he r s , not linguists. A
r e l a t e d noti on i s t h a t i t i s impossible t o c ons t ruc t a th e or y of word
meaning in any scientifically r espe ct abl e form and therefore that
t he r e i s no place f o r ' l i n g u i s t i c l e x i c a l s em ant ic s' .

Let us consider the via b ility of alternative approaches to this

problem, starting wi th a rephrasi ng of Thomason's 'sweeping'
statement. I t can be argued t h a t 'walk' and ' r u n ' differ in meaning
and we require linguistic lexical semantics to provi de semantic
r ep re s e n t a t i o n s for them which should reveal both how they d iffer
from one another (and a number o f r e l a t e d items) and how they are
interrelated. We expect these semantic r e p r e s e n t a t i o n s to be p a rt o f
the l e x i c o n o f a grammatical theor y and a l l the r e l e v a n t i nfor ma ti on
to serve as i nput to l e xi cogr aphy . A dictionary of English is the
output of l ex ic ogr aphy and we expect it to tell us how ' w a l k ' and
'run' d iffer from one another as 'lin g u istic expressions', i.e. as
English words. We do not expect it t o do anything more than t h a t .
After a ll, according to the standard distinction between the
d i c t i o n a r y and the l e xi c on it i s the l a t t e r t h a t i s assumed to be a
'mental lexicon', a reflection of a speaker' s lexical competence.
Li n gu i st s are expected to find out about a speaker' s 'mental
grammar', i.e. t he i n t e r n a l i z e d set o f r u l e s which enable someone t o
understand and speak h i s / h e r language. Lexicographers are not.

The relation between a s peaker 's mental processes and grammatical

operat ions has been a problem o f major concern f o r p s y c h o l i n g u i s t i c s
since the e a r l y s i x t i e s . The goal o f psychological semantics remains
at least to some people. ... .. _ . , ,
clear, A l t i s t o show how language and the world are r e l a t e d to one
another in the human mind" (Johnson-Laird 1 9 8 2 : 7 ) .

In view of the simple fact that language and the world are
necessarily related through the human mind and t h a t no grammatical

- 14 -
model can be d e sc r ibi ng speakers' competence w h i l e a t the same time
being psychologically unreal, the d i s t i n c t i o n between the goals of
l i n g u i s t i c and psychological semantics cannot be very c l e a r - c u t .

This could be m i s i n t e r p r e t e d as implying t h a t t h e r e i s no place f o r

lexical semantics except as p a r t o f p s y c h o ! i n g u i s t i c t h e or y . In f a c t
it should be understood as l e ad i ng to the conclusion that it is
necessary to i n v e s t i g a t e the p o s s i b i l i t y of integrating some o f the
i n s i g h t s o f the philosophy o f language and p s y c h o l i n g u i s t i c s w i t h i n a
lin g u is tic lexical semantic t he o ry .

The majority of relatively r e ce nt work in lexical semantics is

c a r r i e d out w i t h i n the framework o f ' d e f i n i t i o n a l ' systems. Some of
the str onge st a t ta c ks mounted agai ns t d e f i n i t i o n a l systems are based
on t h e i r psychological unreality. It seems d e s i r a b l e at t h i s p oi n t
to consider a s p e c i f i c example in some d e t a i l . This example concerns
the psychological reality o f anal ysing causati ves and d e f i n i n g them
as complex verbs, i.e. breaking them up i n t o a c au s at iv e element and
some o t h e r e le ment(s) also present in corresponding non- ca us at iv es.
for example
Fodor, Garrett, Walker and Parkes (hencef orth FGWP) (1 980 ), Areport
results of tests proving (in their view) that c ausati ves are deep
simplex verbs and t h e r e f o r e undefined. The implication is that no
psychological reality can be claimed for any definitions, since
causati ves are by f a r the ' b e s t ' cases f o r d e f i n i t i o n a l systems.

The method adopted by FGWP f o r t e s t i n g the psychological reality of

anal ysing ' k i l l ' , f o r i n st an ce , as 'cause t o d i e ' can be summed up as
follows. The sentences John bit Mary and John killed Mary are
compared, of which only the latter sentence is considered as
involving a c au sa ti ve verb, and s u b j ec t s ' judgments are sought on
whether it is in the former or the l a t t e r sentence t h a t 'John' and
'Mary' are more " c l o s e l y related". The i m p l i c a t i o n is that if the
l a t t e r sentence does not get a lower score than t he former one ( i . e .
if it is not considered that 'John' and 'Mary' are l es s closely
r e l a t e d as arguments o f ' k i l l ' than o f ' b i t e ' ) , the a n al y s i s o f verbs
i n t o components rece ive s a mortal blow.

- 15 -
Rather expectedl y no correlation was found to exist between
c aus ati ve s and relative complexity in t he specific sense of the
exper iment er s' equation of causativity = compl exi ty = loose link
between the verb's arguments. It is doubt ful whether relative
complexity can be checked through such methods. For i t is hard to
imagine even the f i r m e s t supporter o f d e f i n i t i o n a l systems expecting
such a one-to- one correspondence between l i n g u i s t i c and psychological
phenomena. FGWP's method is r emini scent o f experiments c a r r i e d out
in the sixties with a view to di scov er ing whether derivational
compl exi ty correlated with processing compl ex ity . The hypothesis
tested in those experiments was that the more complex the
transformational derivation o f a sentence is, the more d i f f i c u l t it
would be to produce or comprehend. The results of those
investigations proved in essence t h a t no one-to-one correspondence
should be expected between mental processes and grammatical
operatio ns.

Evidence in favour o f the psychological r e a l i t y o f surface s t r u c t u r e s

is re por te d in FGWP, summing up Levelt's (1970) findings to the
e f fe c t th a t subjects' intuitions r espect t he grouping of words i n t o
surface c o n s t i t u e n t s . A possi bl e expl ana ti on o f s u b j ec t s ' support of
' sta nda rd surface o r d e r ' could be simply t h a t t h i s i s p r e c i s e l y the
or der t o which they are most exposed. I t can be argued t h a t t h i s is
even more so concerning lexical items. I nf or ma ti on 'fixed' and
c onsoli dated in the form o f a s i n g le lexical item (e.g. 'k ill') is
clearly much more immediately accessi bl e in this form than in any
corresponding decomposition (e.g. 'cause t o d ie '). If 'k ill' were
found t o i n vol ve a l o os er link between i t s arguments than 'b ite ',
t h a t would not have been pr oof o f the psychological reality of its
ana ly s is as 'cause to die' either. For 'b ite ' a lso involves a
complex i n t e r n a l make-up. I t might have c o n s t i t u t e d evidence agai nst
the psychological reality of l e x i c a l i z a t i o n , i.e. o f the s al i e n c e of
lexical items as (fixed) units. It is undoubtedly a ppr opr i at e to
seek psychol ogi cal evidence for any linguistic hypothesis set up.
The appropri ateness o f method, i s , however, a serious problem.

Be t h a t as it may, the fact remains that the distinction between

'b asic' and 'complex' concepts assumed in most d e f i n i t i o n a l systems

- 16 -
is not well-established. There is no evidence in f avour of a
hypothesis along the lines: 'c ause -to -die' is l e ss complex than
'k ill' and ' k i l l ' i s more complex than 'd ie ' simply because 'cause'
and 'die' are part o f the definition of 'k ill'. Even i f 'die' is
f u r t h e r analysed i n t o ' s i m p l e r ' u n i t s , the problem w i l l remain o f how
these latter u ni t s (whether they be primitive or not) are
interpreted, i.e. what f i x e s t h e i r extensi ons. Despit e t h e i r m e r i t s ,
which seem t o me t o lie elsewhere (and w i l l be discussed in the
following section), d e f i n i t i o n a l systems have not as y e t provided an
answer t o the o v e r a l l problem o f what r e l a t e s words t o the worl d. We
are s t i l l badly in need o f ext ens ive i n v e s t i g a t i o n t o throw l i g h t on
the process o f c a t e g o r i z a t i o n t h a t converts what Labov (1978) calls
"the continua of the r e al world" into the c at e g o r i e s of the
linguistic system. Unless such conversion processes are better
understood no hi gher level linguistic semantic t he or y can be
expected. The distinction between 'purely linguistic' i nfor ma ti on
and 'knowledge o f the w or l d' is much l ess than a h e l p f u l one.

Consider a s p e c i f i c example provided in illu stratio n of th is famous

distinction by a non-supporter of FS, Mereu (1983). Interestingly
enough the same verb ' w a l k ' i s a t issue, also used as an example by
Thomason (1974) and Cresswell (1978) as mentioned at the beginning of
this s ec t i o n . This time 'walk' is said to have a l e x i c a l meaning
approximately equivalent to "moving in a certain way by means of
legs" and an encyclopaedic meaning, which involves, among other
things, i n fo rm at io n to the e f f e c t t h a t (a) " f i s h cannot w al k" . But
if 'legs' is somehow i ncluded, then r ef ere nc e is i m p l i c i t l y made to
those e n t i t i e s which possess t h i s p r o pe r ty . It is hard t o imagine
that the meaning of 'legs' is arrived at separately from the r e a l
world obj ect s which have l e g s. Therefore?- knowledge o f 'walk' and
'fis h ' i mpl ies knowledge o f ( a ) . So the (a) type o f i n f or m at i on need
not appear e i t h e r in the l e x i c o n or in the d i c t i o n a r y . We are l e f t
e x a c t l y where we were. At one end o f the scale (i.e . no knowledge-
of-the-w orld i n fo rm at io n at a ll) 'walk' and 'run' w ill r e ce i ve
identical semantic r e p r es e n t a t i o n s (as a lr ea dy mentioned in
connection w it h FS). Similarly for 'c raw l' and a l l such change-of-
l o c a t i o n verbs, for clearly if 'legs' can be shown t o l i n k d i r e c t l y
with encyclopaedic information, so can 'contact wi th ground',

- 17 -
presumably, ( o r whatever e ls e w i l l be needed to d i f f e r e n t i a t e between
d i f f e r e n t types o f m ot i o n) . To be c o n s i s t e n t wi th t he 'no knowledge
o f the wor l d' doctrine it is conceivable t h a t even ' mot ion' has t o
di sappear from the semantic specification of 'w alk'; in s h o r t , that
only logical connectives and quantifiers can be handled in a
s c i e n t i f i c a l l y acceptabl e manner.

One more view has to be considered b r i e f l y , which i s advocated by a

number of people working in models of linguistic comprehension,
a rtificia l intelligence and computational linguistics. It entails
elim inating t he distinction between linguistic and encyclopaedic
i n fo rm at io n and r e p l a c i n g the two l e v e l s of a n al y s i s w i t h a unique
one. The under lyi ng assumption i s t h a t world-knowledge i s the f i r s t
and most important source o f i n for ma ti on in decoding the l i n g u i s t i c
message. The 'purely linguistic' contribution of the word is
restricted to s t r i n g s o f sounds and a set o f very simple s y n t a c t i c
rules. Quite c l e a r l y t h e r e i s , in t h i s case, no room f o r a l e x i c a l
level o f processing. Linguistic lexical semantics is again accused
o f adopting an e s s e n t i a l l y ' s t a t i c ' d i c t i o n a r y view.

A final note has to be added concerning the r e l a t i o n o f word-meaning

t o sentence-meaning. The l e x i c o n is sometimes d i s t i n g u i s h e d from the
d i c t i o n a r y by saying t h a t the former i s s ent en c e- or i en t ed w h i le the
latter is word-oriented. The tenet that the semantic status of
individual lexical items is determined by the contribution these
items make to t he meaning o f the sentences in which they appear i s
too wel l known t o r e q u i r e e l a b o r a t i o n . On the ot he r hand, the view
that the meaning o f a sentence is a f un c t i o n of the morphemes it
contains and the way in which those morphemes are syntactically
combined seems equally plausible. The whole issue of whether
sentence-meaning or word-meaning is more basic does not answer any
more questions than i t r a i s e s .

What seems to me a reasonable proposal in relation to word and

sentence meaning is combining the logical form o f a sentence wi th
lexically analysed words which belong to this sentence. The
condi ti ons under which the sentence would be t r u e are to be provided

^(See discussion in chapter 10 of Johnson-Laird 1 983 )

- 18 -
by the truth-theory. This can take care of the contribution of
' l o g i c a l words' and 'se ma nt ic o- gr ammat i ca l' s t r u c t u r e s . I t cannot do
anything more than t h a t . All ot he r ' n o n - l o g i c a l ' l e x i c a l items have
to be analysed separately and d i f f e r e n t l y and t h i s is exactly the
domain o f l i n g u i s t i c lexical semantics. Such items may be analysed
on the basis of the c ondi ti ons of their application to extra-
lin g u istic e n titie s . Actual l i n g u i s t i c expressions and condi ti ons o f
appropriateness o f a p p l i c a t i o n are perhaps the only t hi ngs t o which
we have immediate access.

The result of such analyses w ill not amount to necessary and

s u f f i c i e n t conditions. It w ill be a c t u a l l y argued in the f o l l o w i n g
secti ons that lexical a n al y si s should not have necessary and
sufficient c ondi ti ons as its goal. It can however, cont ai n a lot
more necessary, i.e. i mpor tant, i nf or ma ti on than simple e n t ai lm e nt
r e l a t i o n s can a ll o w f o r .

In any case, whatever is adopted as an a p pr o p r i a t e framework for

lexical a na ly s is cannot be o f the same s tatus as t r u t h condi ti ons
provided for sentences. It may for the most part be language
specific. Comparison o f s i m i l a r l y c a r r i e d out analyses might one day
yield conceptual entities in the form of s u bst ant i ve semantic
universals. This I understand t o be the end product; not the tool to
c ar r y out l e x i c a l a n a l y s i s in t he f i r s t pl ace .

What seems i m pe ra t iv e in order to attempt any lexical semantic

a nal ys is which purports to contribute to an understanding o f what
belongs to linguistic lexical competence is some v i a b l e noti on of
human categorization which can serve as a guide. What is also
necessary i s some t he o ry o f language and the worl d.

The f a c t t h a t no such t h e o r i e s are w i de ly accepted and ready to be

a ppl ied f o r the purposes o f the a n al y si s o f a s p e c i f i c p a r t o f the
vocabulary of a n a tu ra l language does not imply that they do not
e x i s t in a t e s t a b l e form. Lexi cal a na ly s is cannot be postponed u n t i l
such theories are completed. Rather, it can be expected to
c o n t r i b u t e to t h e i r completion by consi deri ng l e x i c a l semantic f a c t s

- 19 -
in detail and p r ov i di ng empi ri ca l c o r r o bo r at i o n and linguistic
preciseness as best i t can.

1 .2 E xi s t i n g t h e o r e t i c a l frameworks: a di scussion

1.2.1 Componential a n a l y s i s : the 'major'shortcomi ngs

Lexi cal semantics has always concentrated on two areas of

investigation: the internal semantic make-up of words and the
semantic relations between them. Componential analysis and field
t he or y have been developed to cover these areas.

Since the early s e v en ti e s, componential analysis (or lexical

decomposi ti on)1 has been accused o f a number o f shortcomings which
cannot be overlooked. In f a c t , these a t ta c ks were meant t o demolish
the whole idea o f semantic components. The e x t en t to which I b e li ev e
these c r i t i c i s m s to have succeeded in t h e i r aim w i l l emerge from a
more or l ess d e t a i l e d discussion o f each one o f them.

1.2.1.1 Nature of features

Abstractness as a c h a r a c t e r i s t i c of meaning components was probably

inherited from two d i f f e r e n t sources: traditional philosophy and de
Saussure's ' l a n g u e ' , but was attacked more s e v er el y than e i t h e r .

Not ic e that abst ractness is a problem in its e lf. Few f i r s t y ear

undergraduates are happy wi th the notion o f the 'phoneme' u n t i l they
get to know what you use i t for; c e r t i f y i n g about something t h a t it
is a b s t r a c t may simply amount t o saying what i t i s not ( l i k e saying
that the phoneme is neither a sound nor a letter, for instance).
Besides t h a t , however, one could clai m t h a t the abstract unit s/he
has set up is a useful t h e o r e t i c a l c onst r uct and proceed to show what
use s/he can make o f i t . I t seems to me beyond doubt t h a t by s e t t i n g
up semantic components one can account in a rather neat way f o r a
number of lexical semantic facts such as synonymy, antonymy,
redundancy, ambiguit y, anomaly, contradiction: they have proved

- 20 -
useful. Theoretical const ruct s are o f course a t ta cke d on what they
cannot r a t h e r than on what they can do. This i s probably a much more
fruitful engagement than di scussing the natur e of the semantic
components proposed by Katz and Fodor (1963) and l a t e r r e de f in e d by
Katz (1972) under the names 'semantic markers' and 'semantic
d is tin g u ish e s '. Yet the c r i t i c i s m o f both versi ons has centred on
two p oi n t s : abst ractness and the d i s t i n c t i o n between markers and
distinguishers.

What seems t o me, however, to be much more i mportant than e i t h e r , is

whether semantic f e a t u r e s can po s si b ly be exhausti ve o f the meaning
o f a word even when they are accompanied by arguments attached to the
p r ed i ca t es and h i e r a r c h i c a l l y structured. Not ice t h a t problems have
cropped up even in the best cases of lexical decomposition, e.g.
'k ill' and ' c a u s e - t o - d i e ' . Although (1) John k i l l e d Mary does e n t a i l
(2) John caused Mary t o d i e , the mere fact that (2) i nvolves two
pr e di ca t es r a t h e r than one gives r i s e to the discrepancy between ( ! ' )
and ( 2 ' ) :

(1 ') *John k i l l e d Mary on Saturday by stabbing her on Fri day

(2') John caused Mary to die on Saturday by stabbing her on
F r i d ay .

To ensure the r e q u i re d synchronicity, time i ndi ce s can be added to

the p r e d i c a t e s . Yet the f a c t remains t h a t b i c o n d i t i o n a l i t y cannot be
expected t o e x i s t between (1) and ( 2 ) ( F o d o r 1970).

Neverthel ess some s o l u t i o n s are a v a i l a b l e :

( a) Lexical decomposition need not amount t o p r ov i di ng necessary and

sufficient c ondi ti ons for a given word to denote; and t h e r e f o r e
f e a t ur e s a r e, in p r a c t i c e , s i m i l a r to meaning p o s t u l a t e s .

(b) Words are 'translated' into a number of components which are

again to be understood as o the r (simpler) words of the same
language. C l e a r l y each language i s e n t i t l e d t o have i t s lexemes

- 21 -
 analysed i n t o terms drawn from i t s own vocabulary ( o r from a more
w id el y understood but culturally akin one) w i t h ou t pretending
that these are, through magic, ele vat ed to the s tatus of a
metalanguage pr oper . Working within the lim its of a specific
language and t he l i m i t a t i o n s of a specific culture and ignori ng
aspirations a t u n i v e r s a l i t y does not imply t h a t one is l e f t wi th
nothing at all as it is of t en suggested by c r i t i c s of lexical
decomposition. With a sufficient number of such analyses at
hand, cross-cultural and i n t e r - l a n g u a g e investigations could be
attempted to yield, at the least, groupings of typological
s ignificance and, at best , theoretical (i.e . metalinguistic)
c onst ruct s arrived at as a result of the interaction of such
analyses and p u r e l y t h e o r e t i c a l c o ns ide r at ions .

( c) Semantic f e a t u r e s can become s u b st a n t i a l if they are adequately

analysed and e xp l ai ne d, irrespective of how e x a c t l y they are
expressed (whether in Engl ish, in Tagalog or in s ym bol s )^ a

good example of this is the extremely interesting attempt by

M iller and Johnson-Laird (1976) to establish cognitively
primitive concepts. This is a way of elim inating the
shortcomings of 'abstraction' and s till being left with a
semantic r e p r e s e n t a t i o n . Such approaches are not to be confused
(as they sometimes ar e) with Katz (1972), for i ns t an c e, where
semantic markers were said to r epr esent 'conceptual components o f
senses' and distinguishers to mark 'purely perceptual
distinctions' w i t ho ut explaining satisfactorily t he content of
e ither.

Lehrer ( 19 74:176) seems to complain that "i n most tr eatment s the

features are le ft unexplained or left for psychol ogi sts or
philosophers to explain". Yet if the components are to acquire
independent substance t h e re are few ot he r a l t e r n a t i v e s . One attempt
at establishing some o f these notori ous semantic primes i nstead of
i n d e f i n i t e l y t h e o r i z i n g about them i s Wierzbicka ( 1 9 7 2 ) .

Wi er zbi cka a t ta c k s the use o f formulae o f symbolic l o g i c in l e x i c a l

semantic representations as not being 'explications', in the sense

- 22 -
that they would require an e xpl anati on in their tu rn. Her own
proposal amounts t o c o n st r uc t in g a paraphrase of a sentence under
ana ly s is which she calls "the semantic representation" of that
sentence by v i r t u e o f i t s being made up o f words taken as primary and
combined in accordance w it h grammatical rules o f what she c a l l s a
"semantic language". The whole const ruct draws e x c l u s i v e l y from her
own i n t r o s p e c t i o n and intuition: "the method is introspection, the
evidence-facts of intuition" ( p . 24). Wi erzbi cka, who a t ta cks the
formulae of symbolic logic as being themselves in need of
e x p l i c a t i o n s , sets h e r s e l f the goal o f e s t a b l i s h i n g primes which w i l l
be expressions which are themselves "impossible to satisfactorily
explicate" ( p . 1 3) . They are expressions in natural language from
which the meanings o f ot he r expressions are b u i l t . She proceeds to
compile a lis t of such ' indefinables' which are supposed to be
adequate t o explicate all u tt e r a n c e s . Apparentl y she assumes t h a t
her formulae are not i n need o f e x p l i c a t i o n simply because they are
drawn from na t ur a l language. In practice, however, the actual
formulae are even more obscure than most o f ' t e c h n i c a l ' metalanguage.
Consider her d e f i n i t i o n o f " x i s moving" :

"x can be thought o f as becoming a p a r t o f d i f f e r e n t pa rts

o f t h a t p a r t o f the world" ( p . 97)

which i s based on A. BogusTawski's idea t h a t movement is "becoming

somewhere". It is not obvious t h a t such d e f i n i t i o n s correspond to
"ideas which everybody can f i n d in hi msel f" (p. 1 5 ) . They may well
correspond to ideas that BogusTawski and Wi erzbi cka 'find in
the ms el ves '. The r e s t o f us do need an e x p l i c a t i o n . So, making an
appeal t o " i n t u i t i v e obviousness" does not seem t o solve t he problem.
What can be l ess counter-intuitive than explicating 's a lt' through
's a lty' and ' e a r s ' through ' h e a r i n g ' , both o f which Wi erzbi cka fi n ds
a ' compell ing temptation'? Notice that even CAUSE which is almost
unanimously regarded as a major c a t e g o r i z i n g element f o r verbs i s not
granted the s t a t u s o f a semantic p r i m i t i v e by Wi e rz bi ck a , because " i t
i s r e l a t e d t o and paraphrasable in terms o f j f " ( p . 17), which i s not
considered a p r i m i t i v e e i t h e r , needless to say. In shor t the problem
is not so much whether we need an e x p l i c a t i o n to understand the
formulae or n o t, as t h a t a lot of de fin itio n s (a) are a r b i t a r y and

- 23 -
 (b) cannot be put to the t e s t . We are thus l e f t w i t h o u t the grounds
for a fruitful di scussion even. Wi erzbi cka herself raises the
problem of "sensual data" objecting to Locke's considering
expressions r e l a t e d to such data i n d e f i n a b l e . Yet her own d e f i n i t i o n
of 'lig h t' is based p r e c i s e l y on such an expressi on: "There is no
light here = This place is such t h a t being in t h i s pl ace one cannot
see" ( p . 19). Identifying semantic features wi th primitives and
seeking p r i m i t i v e s in such ' p h i l o s o p h i c - p h i l o l o g i c a l ' ways seems an
impossible t a s k.

Consider, however, a much b e t t e r founded attempt a t t e s t e d in the work

of M ille r and Johnson-Laird (1976), which shows t h a t an a nal ys is in
terms o f f e a t u r e s or c ondi ti ons can be a useful tool if i t s elements
are pr op e rl y d e f i n e d . 3 M i l l e r and Johnson-Laird are not attempti ng a
thorough exploration of 'cognitive and a f f e c t i v e language' for the
areas they are examining. That could only be the goal o f a mammoth
project. They have at least tried to develop a set of primitives
ivated by th e psy ch o l o g y , of perception and c o n c e p t i o n Jhis is clearly a
safer way to try and establish semantic units (whether they are
called features or conditions) which have some substance than j u s t
ph i l o so p hi z i ng about them. Their specific proposals concerning
'cause' and 'mot ion' are discussed in some d e t a i l elsewhere as they
are particularly relevant to the anal ys is of Modern Greek motion
verbs ( he ncef or th MGMVs). Their form alization does borrow from
symbolic logic, but it is explicable, adequately explicated in the
text its e lf and informative. A brief comparison between their
understanding of causation in terms of ' p e r c ei v ed causes' and
I kegami 's (1969) w i l l h o p e f u l l y show the advantages o f t h e i r approach
over a vague e x p la n at io n (although I seem t o find fault with both
analyses o f CAUSE).

1.2.1.2 Markedness
Markedness i s also i n h e r i t e d from s t r u c t u r a l i s t phonology and u s u al l y
attacked in connection w it h bi n ar y complementary features. The
question i s o f t en r a i s e d as to which member o f an opposi ti on should
take the negation ope rat or and by implication be attributed the
s ta t us o f the l es s ' b a s i c ' one.

*(ibid.:705)

- 24 -
The usual 'lin g u is tic ' evidence brought in support of markedness
consi sts in observing, e.g . t h a t How short is x? presupposes t h a t x
is s h o rt , whereas How t a l i is x? is a n e ut ra l question about x ' s
h e i g ht , from which it is concluded that 'short' is the marked and
'ta ll' the unmarked member o f the op po s it io n. I b e l i e v e t h a t the re
is quite strong linguistic evidence of this type within specific
lexical fields and that it is actually one of the mer it s of the
lexical decomposition and lexical field theories to have drawn
a t t e n t i o n to t h i s phenomenon in connection wi th semantics. I t can be
extended t o apply to d i f f e r e n t inter-lexeme relations and shown to
pl ay a role in the o r g a n i z a t i o n o f semantic s u b f i e l d s . This poi nt
w ill be taken up whenever it arises in the course of the present
analysis. Examples from Modern Greek (hencef orth MG) compounds and
semantic s im ilarity tests on MGMVs po i nt to the direction of
structures in p a i r s and a t the same time the r e l a t i v e 'p rio rity ' of
one o f the two members. There are a number o f different ways to
account f o r t h i s r e l a t i v e p r i o r i t y . Markedness i s probably the most
general of them and the best e s t ab l i s he d one. So rather than
consi deri ng i t a problem s p e c i f i c t o l e x i c a l decomposition, i t should
be understood as a phenomenon o f wide a p p l i c a t i o n .

1.2.1.3 A t o mi c it y and u n i v e r s a l i t y o f f e a tu r es
Features are supposed to stand f o r atomic concepts, i.e. unanalysable
units . This is not to be understood as necessarily implying the
individual and separate e xi s te nce o f e n t i t i e s ; in f a c t most work o f
the componential a n al y s i s type was actually in keeping wi th the
genuine structuralist spirit of emphasizing the interdependence of
e n t i t i e s and t h i s i s , in my view, one o f the most e s s e n t i a l meri ts of
any such practice. At the same t ime , though, features are also
supposed to belong t o a u ni ver sa l set from which i n d i v i d u a l languages
select and draw, making different combinations in the process of
lexicalization. This a l le g ed universality of sense components has
been easily (and for obvious reasons) attacked ever since it
appeared. In f a c t , one method o f br i ngi ng the whole idea o f l e x i c a l
decomposition into d i s re p ut e is through c lai mi ng that in these
aspirations to universality lay most of the attraction of
componential analysis (see Lyons 1981). I have strong doubts about

- 25 -
this c la i m but very few doubts that universality and atomicity
c o n s t i t u t e the s or es t poi nts in semantic f e a t u r e t h e o r i e s . This is
not to be understood as implying t h a t the search f o r d i f f e r e n t types
of semantic universals is doomed to failure. Different thematic
relations as expl ored by F i l l m o r e as wel l as Gruber and Oackendoff
might be thought of as a step in e s t a b l i s h i n g semantic u n i v e r s a l s .
Parallel to t he se, investigations in quite specific areas of the
vocabulary such as B e r l i n and Kay's (1969) well-known study o f col our
terms, Lehrer's (1974) cooking ter mi nol ogy, Greenberg's (1963)
statistical universals are v er y interesting and quite successful
attempts in r e s t r i c t e d domains, although, in a way, following the
American Ant hr opol ogical tradition. Universal semantic features
might some day grow out o f the r e s u l t s o f such i n v e s t i g a t i o n s .

Whether G. Frege i s r i g h t l y or wrongly 'accused' o f being respons ibl e

for the 'com positionality principle' I am in no p o s i t i o n t o judge.
What I do know, however, i s t h a t ' the p a r a l l e l i s m between decomposing
chemical substances and words, i f taken l i t e r a l l y (as i t o f t e n seems
to be) i s u n s u cc es sf u l . 4

There are, however, some understandings of atomicity which are

compatible w it h componential a n al y s i s and much less objectionable
than the noti on o f u n i v e r s a l , i n h e r e n t l y unanalysable u n i t s .

F i r s t of a ll, as i t is ext remel y d i f f i c u l t , if at a l l po s si b le , to

separate ' p r i m i t i v e ' from ' n o n - p r i m i t i v e ' concepts and as t h e r e i s no
agreement y e t on the c rite ria for this distinction, 'un a na ly s ab l e'
could be modi fi ed to imply: 'not further analysed for t he time
being', i.e. at the present state of knowledge. Now, whether t h i s
knowledge is to be f u r t h e r e d through a better understanding of t he
r u l e s governing perception, or a wi der and, a t the same t i m e, more
d e t a i l e d e l a b o r a t i o n o f f u n c t i o n a l p r o p e r t i e s o f o b j e ct s or , in f a c t ,
through the development of an adequate theor y of t he link between
lexical competence and l e x i c a l performance w i l l every time depend on
the natur e o f t he domain under a n al y s i s and the basic phi l osop hi ca l
stand o f the a n a l y s t .

- 26 -
Secondly, if one understands the purpose of the a n al y s i s to be
br ingi ng out the relations between c l o s e l y connected l e x i c a l items
through an investigation of their inte rnal structure, i.e.
concentrate on the interdependence o f l e x i c a l u ni t s ( i n the standard
structuralist tra d itio n ), the question o f a n a l y s a b i l i t y in terms o f
an i nhe rent pr ope rt y o f f e a t u r e s does not even a r i s e . One posi ts as
components those p a r t i c u l a r items which are c o n t r a s t i v e and help in
establishing a n e at , elegant, economic and sometimes i n fo r m a t i v e
paradigm. Naturally, no notice is taken of the nature of the
relation between one's c onst ruct s and e i t h e r the world or the human
mind; therefore 'informative' here is to be understood in this
r e s t r i c t e d sense, i.e. burdened w it h these l i m i t a t i o n s .

So, Lounsbury's (1964) features used for the a n al y s i s of kinshi p

terms are the semantic dimensions of the particular field he
investigates and not f u r t h e r analysed. In t h a t sense they can also
be understood as atomic, although Lyons ( 1 9 8 1 ) , who equates ' a t o m i c '
wi th 'separate and i ndependent' , cont ra st s the noti on of atomicity
wi th s t r u c t u r a l i s t interdependence.

Cos er iu' s 'primary' vocabulary - to tur n to European s t r u c t u r a l i s m -

c onsi sts o f words which "do not imply ot he r words, but correspond to
immediate experience" (Coseriu and Geckeler 1981:56). It is not,
however, in the l e a s t obvious why the f r u i t of a tre e (e.g. 'pomme'
offered as an i nstance o f a 'primary' word) i s more accessi bl e to
immediate experience than the t r e e i t s e l f ( i . e . 'pommier' given as an
example o f a 'secondary' word). Morphological derivations are one
thing, immediate experi ence i s another. S t r u c t u r a l i s m i s a t i t s best
when i t concentrates on f u n c t i o n a l linguistic opposi ti ons and drops
i t s claims as t o the r e l a t i o n s h i p between two d i f f e r e n t systems, i . e .
language and the w o rl d .

Yet , it must be admitted that the commonest understanding of

'primary' is probably the one implying (if not d irectly stating)
'psychologically and logically independent and corresponding to a
primary conceptual u n it', t h e r e f o r e i n h e r e n t l y unanalysable and thus

- 27 -
atomic. This i s p r e c i s e l y the kind o f a t o m i c i t y which i s completely
open to c r i t i c i s m .

1.2.2 Componential a n a l y s i s and semantic f i e l d t h e o r i e s viewed

as s t r u c t u r a l i s t t h e o r i e s

I have so f a r concentrated on l e x i c a l analyses based on components

but have on purpose also r e f e r r e d to semantic f i e l d t heor y as I do
not f e e l t h a t the two should be kept separ ate. To s t a r t w i t h , most
analyses of semantic fields are carried out wi th the aid of
categorizing units more o ft e n than not in the form of components
(i.e . both the American Ant hropol ogical p r a c t i c e o f anal ysing kinshi p
terms using components like ' ma le' and the European stru cturalist
t r a d i t i o n o f using elements l i k e 'arms' or 'back' when analysing the
s u b -fie ld of ' s e a t s ' ) .

Secondly, to do j u s t i c e to both types o f a n a l y s i s , it i s important to

see how they complement each o t he r . Any kind of componential
a na ly s is w i l l make use o f r e c u r r e n t sense components. It therefore
provides a tool f o r d e f i n i n g d i f f e r e n t conceptual areas on the basis
o f such components.

Thirdly, it is my view t h a t both approaches have been extremely

harshl y criticized , often on t he same grounds and t o a po i nt where
t h e i r me r it s have been overlooked.

What has a lr ea dy been said in connection wi th uni ver sal ism should not
be understood as implying that structuralist linguistic theories
cannot c o n t r i b u t e t o the est abl ishment o f u n i v e r s a l s . For i nstance,
the f a c t t h a t d i f f e r e n t languages t r e a t 'space' d i f f e r e n t l y and t h a t
this w ill necessarily gi ve rise to different categorizations as
results of the application of semantic field t he or y does not
invalidate the actual analyses. Successful analyses of similar
fields in d i f f e r e n t languages can l e ad , a t the l e a s t , to typological
c a t e g o r i z a t i o n s o f languages (as a lr e ad y mentioned) and, at best , to
a better understanding of how ' space' is c onc ep tu al i ze d, provided

- 28 -
issues of substance a lr eady discussed are a lso taken into
consideration. A dm it t edl y the most successful anal yses, among those
conducted w i t h i n the framework o f such methods, are probably the ones
d e al i n g wi th 'functional' areas of the vocabulary (for i nstance
prepositions) and this is to be expected in structuralism , where
're la tio n a l' i s p r a c t i c a l l y equated wi th ' e s s e n t i a l ' , i.e. where each
item i s i d e n t i f i e d on the basis o f i t s r e l a t i o n s wi th ot h e r s . There
is, however, no reason to suppose t h a t s t r u c t u r a l i s m is in p r i n c i p l e
prohibitive as to c on si d er a ti o ns o f substance. Besides, it remains
to be seen whether i t s major premise w it h respect t o word meanings,
namely t h a t they are not independent of one another, is basically
wrong as it is sometimes claimed (e.g. Verschueren 1 98 1: 329 ). It
seems t h a t t h e r e is quite strong evidence from language a c q u i s i t i o n
that the meaning o f one word i s actually learned by simultaneously
l e a r n i n g the meaning o f ot he r words. Moreover, it is q u i t e evi dent
that semantic components i n vol ve dimensions of contrast and it is
d iffic u lt to believe that these c ont ra st s have no bearing on the
issue of conceptualization. This is a major issue which r equi re s
careful c o n si d e r at i o n and w i l l be taken up again in sec ti on 1.3 in
connection wi th protot ype and ste reot ype approaches t o meaning.

Finally, one of the most severe a t ta cks mounted a g ai n st lexical

decomposition and semantic f i e l d theories, namely t h a t they are not
exhausti ve anal yses, has to be seen in the l i g h t o f what a l t e r n a t i v e
s o lu t i o n s have to o f f e r in t h i s respect both from the t h e o r e t i c a l and
the practical p oi n t of view. So, meaning p os t ul a te s which are
supposed to' have fewer t h e o r e t i c a l problems, y i e l d at l e a s t e q ua l l y
non-exhaustive anal yses, w hi le t h e i r proper domain i s s t i l l uncl ear .

1.2.3 The empi ri ca l v a l i d i t y o f componential a na ly s is

A number o f arguments have been put fo rth a gai nst t he validity of

componential a n al y s i s as a practical tool for the study of word
meaning. A b r i e f c o n si d er a ti o n o f the arguments in Lyons (1981) is
in or der here, on the basis of which Lyons concludes that the
e mpi ri ca l valid ity of componential a n al y s is is "more apparent than
real" (p .83). Lyons seems to o bj e c t t o anal ysing ' boy ' as 'human';

- 29 -
he a c t u a l l y s t a t e s t h a t this pr op e rt y i s not an e s s e n t i a l component
of 'boy', since "the male offspring of the gods (e.g. Cupid) are
regularly described as boys ... but ... they are not said to be
human" ( p . 8 4 ) . However, i n tr oduc ing i n t o such a discussion the names
f o r the sons and daughters o f Greek gods does l i t t l e more than touch
on the important ( p h i l o s o p h i c a l ) problem o f which c h a r a c t e r i s t i c s of
an e n t i t y are ' e s s e n t i a l ' ; this issue i s discussed in some d e t a i l in
1 .3 in connection w it h Putnam's stereot ypes which seem to me to
c o n t r i b u t e a l o t to i t s r e s o l u t i o n . I t has no be ar i ng, however, on a
di scussion o f the e mp i r i c a l v a l i d i t y o f componential analysis. Greek
gods were 'humani zed' , i.e. conceived of as 'human', sharing wi th
o r d i n a r y human beings a number o f ' e s s e n t i a l ' properties; sharing in
fact a lot of their 'su p e rficia l' characteristics, which (as w ill
become obvious from the di scussion of s te r eo t yp e s) are the most
'essential' in deter mini ng the meaning of words. Some of their
' de e p e r ' characteristics, e.g. their immortality, were of course
speci al to them, but they were not enough t o 'de-humanize' them in
ot he r r e spe ct s. Hence t h e i r male o f f s p r i n g would be c a l l e d 'boys'
f o r e x a c t l y the same reasons t h a t the male o f f s p r i n g o f non-gods were
called 'boys'.5 To my mind, such consi de ra ti ons prove r a t h e r than
di sprove that ' boy ' i m pl ie s 'human'. Lyons i m pl ie s that meaning
p o s tu l a t e s avoid t he problems o f componential a n al y s i s but he does
not demonstrate a s a t i s f a c t o r y a n al y s i s o f 'boy' in meaning p os t u l at e
terms which could do away wi th 'human' and ' n o n - a d u l t ' and s t i l l be
informative.

From a theoretical point of view meaning p os t u l a t e s inherit the

problems o f semantic e n t ai l m en t in g e n e r a l , the exact nature o f which
we are s t i l l f a r from having understood. So the question seems to be
r a t h e r whether i t makes sense t o l ook f o r exhausti ve d e f i n i t i o n s and
also whether we should expect the c o m p o n e n t s/ f e at u r e s/ c on di t i o ns /
e n t ai l m en ts we p o s i t to be ' s i m p l e r ' than the terms we use them to
define. The answer to both seems to be n e g a t i v e .

- 30 -
1.2.4 Componential a n al y s i s and semantic f i e l d t h e o r i e s :
the ' mi nor ' shortcomings

I have tried so f a r to argue agai ns t the attitude of c ritic izin g

componential a n al y s i s and semantic f i e l d theory too s e v e r e l y , wi thout
r e a l l y s u b s t i t u t i n g anything t h e o r e t i c a l l y c l e a r and i n d i s p u t a b l e in
their p l a ce . It is, however, my view t h a t both types of a nal ysis
s u f f e r from shortcomings ot he r than the ones I have a lr ea dy discussed
and which are considered 'major' shortcomings in the literature.
Before turning to th e se , it seems worth consi deri ng some ' mi nor '
questions such as the d e l i m i t a t i o n o f a semantic f i e l d , formalization
and generality of application in connection wi th semantic field
t he or y.

The d ifficu lty of delimiting a semantic field seems to have been

over-exaggerated as a problem. The question o f whether i t i s in f a c t
possi bl e to determine the boundaries o f a given l e x i c a l area has been
repeatedly put forth. A partial answer to this question is the
exi st ence (in fact or in p r i n c i p l e ) o f crossing f i e l d s , open-ended
fields, fields with sets of items which bear paradi gmati c or
syntagmatic r e l a t i o n s t o one another. There is nothing in the theory
i t s e l f which prevents one from adopting the method best f i t t e d to the
r e l a t i o n s to which one wants to draw a t t e n t i o n . I t h e r e f o r e consider
the m a tt e r a pu re l y p r a c t i c a l one and the f l e x i b i l i t y o f the theory
one o f i t s m e r i t s . In the present a n a l y s i s , an example o f a possible
solution i s provided in connection wi th a p a r t i c u l a r f i e l d , and t h i s
i s done wi th the understanding t h a t one could t h i n k o f a number o f
different ways o f approaching exactly the same problem w i t h i n the
framework of the same theory. That semantic fields are not
necessarily closed and w e l l - d e f i n e d sets (as European s t r u c t u r a l i s m
viewed them) need not be a shortcoming o f the theor y i t s e l f as some
semantic f i e l d t h e o r i s t s consi der ( e . g . Lehrer 1974). I t i s probably
t o be expected as a r e f l e c t i o n o f the nature o f na t u r a l language. In
my view, it only becomes a serious handicap if the goal is to
adequately f o r m a l i z e the theor y so t h a t i t could serve as a component
o f a formal grammar.

- 31 -
As regards the absence o f such a f o r m a l i z a t i o n i t seems r a t h e r naive
to consider i t a mere 'omi ssi on' (going along wi th Lehrer 1974). By
lea vi ng the sentence out o f i t s perspe ct ive ( a t l e a s t in i t s so f a r
e x i s t i n g forms) semantic f i e l d theor y probably cuts i t s e l f away from
the possibility of having its analyses 'translated ' into explicit
formulae of the kind r e q u i re d by formal grammar. It seems to me,
however, that the issue of form alization has wrongly been given
p riority (in c omparati vel y r ece nt year s) over the logically prior
issue o f adequacy. Formal grammar cannot scare away f u zz ine ss if it
is part of the nature of human language. A number o f alternative
theories, a lr ea dy hi nt ed at, seem t o ignore a lot of intuitively
important lexical i n fo rm at io n for the sake of 'fo rm a liz a b ility '.
Paradoxical as i t may sound, I really believe (and w i l l argue when
discussing 'adequacy') that European structuralism has also lost a
g re a t deal in d e s c r i p t i v e adequacy and i nformati veness by imposing a
rather rigid schema on the data it analysed and sometimes drowning
i t s e l f in a sea o f ' t e r m i n o l o g i c a l ' sub-distinctions.

I t can be argued t h a t f i e l d t h e o r i e s have always depended on types o f

lexical material which l e n t themselves by t h e i r very nature to such
analyses. Whether a l l meanings should or could be analysed in the
same terms i s o f course d e ba t ab l e. Yet i f a theor y i s b u i l t on the
p e c u l i a r i t i e s o f very few domains i t may not be general enough in any
i n t e r e s t i n g way. F i e l d t h e o r i e s have produced successful r e s u l t s in
the anal ys is o f k i ns hi p terms and personal pronouns, f o r i nstance; in
ge n er a l , sets of words that contrast paradigmatically and can be
shown to divide what is usually called 'conceptual spac e' . It is
probably at its best in those cases where a small number of
components is s u f f i c i e n t to c o n t r a s t a l a r g e number o f i tems. I t is,
however, the case t h a t not a l l vocabulary i s 'simple' in the sense
that it can be analysed only in structural terms. It has been
pointed out , for i n s t an c e, t h a t notions such as promising, ordering
and the l i k e cannot be t r a n s l a t e d i n t o l e x i c a l decomposition formulae
(see Verschueren 1981:324) because such formulae are inherently
unable t o capture the i d io sy n c r a c i e s o f speech act verbs. Sim ilarly
M i l l e r and Johnson-Laird (1976) argue t h a t important concepts such as
PERSON do not have a coherent lexical field associ ated w i t h them.
This latter criticism is probably e a s i e r t o handle than the former

- 32 -
one. The domains i n t o which 'conceptual space' i s d i v id e d up are not
given and 'per son' might not constitute such a domain. M iller and
Johnson-Laird put however, SPACE i n t o the same basket as not mapped
onto an "intuitively coherent lexical field" ( p . 375). But ' space'
can be e a s i l y understood as a cover-t erm applicable to very l a rg e
nets i n c ludi ng different subsets (e.g. direction, location, motion)
some o f which o v er l a p; these are again s p l i t i n t o areas l e x i c a l i z e d
as s p a t i a l locatives (PPs and Advs f o r i n s t a n c e ) , motion verbs, e t c .
I n t e r l o c k i n g f i e l d s can c o n s t i t u t e q u i t e a coherent whole. I t is not
therefore immediately obvious that ' space' does not constitute an
'in tu itiv e ly c oher ent' domain. Besides, it can be no acci dent t h a t
p a r t i c u l a r sub-areas belonging to the general notion o f 'space' have
been analysed quite s u c c e ss f u l l y wi th semantic field and lexical
decomposition t echni ques.

P r e c i s e l y f o r these reasons i t i s i n t e r e s t i n g to i n v e s t i g a t e what an

alternative approach to word meaning might offer in such an area.
The direction in which to l ook for an alternative can only be
dictated by the prospect of gai ning in descriptive adequacy and
i n for ma ti ve ne ss, as a l re a dy s t a t ed .

1.2.5 S t r u c t u r a l i s m , d e s c r i p t i v e adequacy, and the case o f

motion verbs

It has been pointed out in r ecent years that a number o f lexical

semantic f a c t s are not and cannot be accounted f o r by any ' c h e c k l i s t '
semantic theor y (Fillmore 1975, 1978) but might be handled by
a l t e r n a t i v e approaches commonly r e f e r r e d t o as ' p r o t o t y p e ' ones. The
main t h e o r e t i c a l assumptions o f Prototype and Stereotype t h e o r i e s and
their possi bl e a p p l i c a t i o n to MGMVs are discussed s e p a r a t e l y . Some
poi nts have to be made here, however, with the specific aim of
comparing Prototype t heor y to componential a na ly s is and semantic
field theories (which are the sor t of checklist approaches I am
discussing).

- 33 -
Most componential analyses seem t o asc ri be t o ( o r a t l e a s t a spi re to
fu lfillin g ) the 'minimal definition principle' (Bendix 1 966), which
a lso i mpl ie s what we could c a l l 'maximal generalization' as another
goal o f the ideal definition. The r e s u l t o f such a premise i s t h a t
the d e f i n i t i o n o f a term ( a) accounts f o r a l l pos si b le instances o f
this term and (b) contains precisely those conditions/features/
components which are necessary and s u f f i c i e n t ^ to d i s t i n g u i s h i t from
every o t he r term in the l e x i c o n . Anything e x t r a would be a case of
what we could call 'stru c tu ra lis t redundancy'. It would be
interesting, however, to see if the application of this 'doubl e'
principle has succeeded in practice. The two extreme alternatives
are clear: if it has, we have an e x c e l l e n t method of arriving at
ext reme ly neat and economic d e s c r i p t i o n s ; if it has not , we could
suspect something being wrong wi th the principle its elf, i.e. not
corresponding to the f a c t s o f n a tu ra l language.

There is evidence that, at least for certain types of lexical

m aterial, e.g. speech act verbs, lexical decomposition formulae are
so incomplete as d e s c r i p t i o n s that far from being in a position to
make e x p l i c i t the f u l l meaning o f the terms they d e f i n e , they cannot
even d i f f e r e n t i a t e between related items. Consider, for i nstance,
Verschueren's (1981) examples f o r ' a r g u e' and ' s t a t e ' . His clai m is
that lexical decomposition formulae would be identical for these
verbs (hencef orth Vs) which are c l e a r l y non-synonymous:

s t a t e ( x , y , P ) = S A Y ( x , y , S e ) AI NTEND(x,CAUSE([SAY(x,y,Se)],
COME ABOUT (KN0W(y,P)) ) )

argue ( x , y , P ) = S A Y ( x , y , S e ) AI NTEND(x,CAUSE([SAY(x,y,Se)],
COME ABOUT (KN0W(y,P)) ) )

A lternatively all t he basic components o f a l l speech act Vs would be

c o n f l a t e d i n t o a s i n g l e formula:

S A V ( x , y , ( P ) ) = S A Y ( x , y , S e ) AINTEND(x,CAUSE( [ S A Y ( x , y , S e ) ] ,
COME ABOUT (ACCEPT(y,SA') ) ) )
(Verschueren 1981:325)

- 34 -
Notice also t h a t o t h e r complex Vs, e . g . verbs o f c o g n i t i o n , are oft en
analysed w it h the help of c ondi ti ons rather than componential
formulae, e.g . Lehrer's (1974) 'belief-predicates' or M iller and
J ohnson-Lai rd' s (1976) Vs of communication (among others). The
c ondi ti ons used l ook very much like 'felicity/ap propriateness'
condi ti ons of protot ype semantics in form, but they are in fact
ent ai lme nt s - either e xp lic itly ( Le hr e r) or im plicitly (my
understanding o f M i l l e r and J o h n s o n - L a i r d ) .

Semantic i n fo r ma t io n which cannot be reduced to 'g eneral' and

'simple' one-word features cr ea te s seri ous problems for checklist
theories. The notor ious distinction between 'ma rke rs ' and
'distinguishes' which has convinced very few people as to its
theoretical validity could be seen as an attempt to accommodate ( i n
the form o f d i s t i n g u i s h e r s ) such ' u n s i m p l i f i a b l e ' m a t e r i a l . ^

In shor t, when complex material is i nv olved, either the 'minimal

definition-maximal generalization' principle is abandoned or
d e s c r i p t i v e adequacy. For even w i t h i n domains which g i ve evidence o f
i nh e re nt structuring, i n f or m a t i o n specific to very few items cannot
be i ncluded in a rigid structure wi thout creating 'structuralist
redundancy'. Examples o f analyses exhibiting these c h a r a c t e r i s t i c s
w ill be given in what follows. The main po i nt is that the
combination o f t h i s principle and the 'neat' f o r m a l i z a t i o n goal have
probably done more harm than t he 'main shortcomings' discussed in
this section. Thi s combination has weakened consi de ra bl y the
d e s c r i p t i v e power o f such anal yses.

Contrary to c h e c k l i s t t h e o r i e s which seek t o e s t a b l i s h safe c r i t e r i a

that would guarantee membership in a cat egor y (necessary and
sufficient conditions), Prototype theories seek c riteria
(a ppr opr iateness c o n d i t i o n s ) f o r i d e n t i f y i n g the p r o t o t y p i c a l (ideal/
best ) i nstance o f a cat egor y. This means t h a t the boundaries of a
category are not defined; something can be a clear case of this
category (prototypical i n st an ce ) if all the c ondi ti ons are met; if
only some o f them are satisfied the term may s t i l l be a p p l i c a b l e ,
provided the deviation from the focal p oi n t (prototypical case) is

- 35 -
not too g r e a t . Hence phenomena o f indeterminacy and g r adat io n which
are characteristic of na tur al language can, in principle, be
accounted f o r in terms o f degree o f membership.^

Before engaging in a comparison and an e va l u a t i o n o f the important

theoretical assumptions o f the two types o f approach, I would l i k e to
compare in detail actual pieces of a n al y s i s of checklist and
'prototype' descriptions respectively. The exact st at us or nature o f
the components/conditions plays no role in this particular
comparison, so f o r pr esent purposes the terms 'features-conditions-
components-entailments' w i l l be used as e q u i v a l e n t .

The aim o f t h i s 'exercise' is to show:

1. that prototypical (specific) and general c ondi ti ons are

i n t e rm in g le d in c h e c k l i s t (and some p r o t o t y p i c a l ) d e s c r i p t i o n s ;

2. that 'lin g u is tic ' and 'extra-lin g u is tic ' material is e q u al l y

i n te r m in g l e d ;

3. that t he re are differences not only in kind but also in

complexity between the various components appearing in
componential formulae ( e . g . WITH used in the same formula as ATG
which stands f o r 'always touching g r ou nd ' ) ;

4. that as a result of all these, componential t h e o r i e s cannot

fu lfil the principles they set themselves to f u l f i l , even for
domains in which t hey have produced fa irly successful ( i n the
sense of illuminating) analyses, such as the domain of motion
verbs (hencef or th MVs).

The s t a r t i n g point of th is comparison i s the c l a s s i c example o f an

anti-checklist approach, Fillm ore's (1978) note on the semantics o f
'clim b'. The tentativedefinition provided i ncludes: (a) a
clambering component: locomotory act ion o f movable body pa rt s and (b)
an ascending component. It is hard to come across a less specific

- 36 -
( o r a more g e n er a l ) d e f i n i t i o n f o r ' c l i m b 7, c o nt r a r y t o e xpe ct at ions .
A standard l e x i c a l decomposition anal ysis o f t he term would probably
i nclude additional components like ' intentionality' of the act ion
('agent' or 'actio n ') and a specification of the medium: 'earth/
ground/supporting s u r f a c e ' ; all o f these I would have thought q u i t e
p r o t o t y p i c a l too. Consider the f e a tu re s proposed by act ual c h e c k l i s t
analyses f o r ' c l i m b ' :

Wi erzbi cka (1972) is a good case because she i s a f e r v e n t proponent

o f the 'minimal definition' principle. Although she does not o f f e r
an a n al y s i s o f 'c lim b' itself, one can r a t h e r safely c onst ruct one
w i t h i n the framework she uses on the basis o f her d e f i n i t i o n s f o r a
number o f r e l a t e d motion verbs such as ' c r e e p ' , 'crawl' and 'moving
up' which she d e fi ne s as "becoming f u r t h e r from the Earth" ( p . 104).
Her proposal for 'creep' i nvolves movement of an a g ent ' s 'b e lly ',
w h i le t h a t f o r ' c r a w l ' i ncludes movement of 'arms' and ' l e g s ' . So, a
pos si bl e (restructured) a na ly s is o f : A is cli mbi ng up x would be: "A

causes movements of his arms and legs which cause his body to be
becoming supported by f u r t h e r p a r t s of x and f u r t h e r from the Earth",
I have hi nte d at the shortcomings of the actual formalization
elsewhere. The main p oi n t here i s to noti ce t h a t , in p r a c t i c e , such
a definition is at least just as 'prototypical' as Fillm ore's: in
actual fact, it i s even more so, as i t involves e x p l i c i t re fe re nc e to
human body par ts and the ' i n t e n t i o n a l ' component.

M i l l e r and Johnson-Laird (1976) o f f e r a d e f i n i t i o n f o r ' c l i m b ' o f the

form: ACT (x,(UP(TRAVEL)) ( x , y ) where ACT (x) stands for
CAUSE(INTEND(x)) , i.e. the ' i n t e n t i o n a l ' component and TRAVEL stands
roughly for the 'change-of-locatio n' component. Leaving aside the
f o r m a l i z a t i o n f o r the moment, observe the i n c l u s i o n o f ' i n t e n t i o n ' as
well as the specification of 'ascending' as another c o nd i t i on for
'clim b'. Now M i l l e r and Oohnson-Laird do r e a l i z e t h a t you can 'c li mb
down'; y e t their 'general' definition includes UP (which leads once
more toward the p r o t o t y p i c a l understanding o f ' c l i m b ' ) . They choose
to explain this in terms of 'climb' being "used with a greater
variety of directional m o di f i e rs " ( p . 552) as compared t o 'ris e ', for
i ns tance, which cannot collocate with 'down'. Their conclusion is
that "...th ere are some verbs whose i ncorporated directions can be

- 37 -
ov er r ul ed by e x p l i c i t expressions to the c o n t r a r y and o the r verbs f o r
which any a d d i t i o n a l specification of direction must be c o n si s t e nt
wi th the i nc or por at ed d i r e c t i o n " ( p . 5 5 3 ). This complete absence o f
e xp la na ti on for the facts seems to be the cost of lim iting the
semantic s p e c i f i c a t i o n to m e t a l i n g u i s t i c terms.

To f u l l y a p pr ec i at e the problem, one has t o bear in mind t h a t M i l l e r

and Johnson-Laird are actually comparing 'clim b' to 'ris e ' and
consider t h a t the main d i f f e r e n c e between them is t he c o ndi ti on of
'in te n tio n a lity'. The actual 'manner' o f motion which i s present in
'clim b' and absent in 'ris e ' is not linked to t he c o nd i ti o n of
in tentionality in their description, it is not discussed at a ll,
presumably because for the definition to be 'g eneral' enough no
' c la mbe r ing' c o n d i t i o n can be i ncluded, since a s nai l can be said to
' c l i m b up a f l a g p o l e ' , to use F i l l m o r e ' s example.

Alternatively it cannot be mentioned because 'manner' cannot be

e a s i l y reduced t o ' s i m p l e ' one-word f e a t u r e s . It is c l e a r , however,
t h a t whether the c o n di t i o n of i n t e n t i o n a l i t y is present or absent in
'ris e ', the verb w i l l s till be a p p l i c a b l e (and quite lite ra lly be
used t o o ) , provided the c o n di t i o n o f 'upward' motion is kept (e.g.
"he rose to hi s f e e t " ) . In or de r to i d e n t i f y the nuc l ea r sense of a
term one must have access to removability of c ondi ti ons and
weightings o f i n d i v i d u a l c rite ria . I t i s q u i t e p os si bl e t h a t even i f
the nucl ear sense i s understood as the maximally general one (i.e.
corresponds to the checklist ideal), the way to arrive at it is
through weighing di s ta n ce from focal points. The c o n d i t i o n ( s ) which
's urv iv e ', in the sense t h a t t hey prove unremovable, are probably the
ones which have t o be included in the minimal definition. But if
t h e r e are only two candidates (say ' c l am be r in g' and ' a s c e n d i n g ' ) and
they both prove removable, the checklist definition has an
u nf or t unat e case: it w ill either contai n nothi ng, or contai n ,a
d i s j u n c t i o n or p o s i t two d i f f e r e n t e n t r i e s ( c l i m b j and c li mb 2 ) . None
of these seems too good. The e mpi ri cal validity of prototype
semantics seems t o l i e mainly in its a b i l i t y t o provi de a framework
w i t h i n which one t r i e s t o e s t a b l i s h the norm ( p r o t o t y p i c a l case) f o r
a cat egor y, and account f o r acceptable d e v i a t i o n s from i t . These are

- 38 -
at least useful tools in explaining, rather than just stating,
lexical semantic f a c t s .

Fillmore's attempt at anal ysing ' c l i m b ' i s probably not the best
example of prot ot ype analysis, but i t cannot lead to the
im possibilities ju s t d i s c u s s e d . ^ So, if a monkey can ' c l i m b down a
f l a g p o l e ' w h i le a s n ai l can only ' c li mb up' one, F i l l m o r e can e xpl ai n
the a c c e p t a b i l i t y o f the former instance in terms o f the presence of
the ' c l am b er i n g' c o n di t i o n de s p i t e the absence of t he 'ascendi ng'
one, and the unacceptability of *A s nai l is cli mbi ng down the
flagpole in terms o f the absence of both p r o t o t y p i c a l c ondi ti ons -
i.e. ' ascendi ng' and ' c l am be r in g' - which results in this last
instance of 'c lim b' being too far removed from t he f ocal po i nt
(comprising both c o n d i t i o n s ) .

Two thi ngs are absent from this sketchy presentation of a

'prototypical' analysis of a term: the importance of obta ini ng
speakers' judgments concerning p r o t o t y p i c a l p r o p e r t i e s o f items and a
di scussion of the relative importance of and the relation between
conditions. As an i l l u s t r a t i o n o f the l a s t p o i n t consider the case
of 'clim b', 'run' and 'w alk', on the one hand, versus ' r i s e ' on the
other. I have al re ady hi nted at t he possibility that the actual
'manner' of motion of the former set is closely linked to the
c o nd i t i on of ' i n t e n t i o n a l i t y ' , w h i le f o r ' r i s e ' , which i s uns peci fi ed
as t o 'manner ', the question o f i n t e n t i o n a l i t y i s sepa ra te . A third
case i s e xe m p l i f i e d by ' s l i p ' (as in He sli pped and f e l l down) which
seems t o e x h i b i t a l i n k between 'manner' and ' i n t e n t i o n a l i t y ' in the
opposite direction (absence of the pr oper ty in the prototypical
understanding of the term). There is evidence from tests that
speakers a ssoci ate (such) discrete properties w it h lexical items,
although t h e i r actual nature and the e xt ent to which they are aware
o f them i s an open quest ion. So, if prototype/stereotype analyses
can be constructed so as t o y i e l d a fin ite list of properties fo r a
lexeme (which i s doubtful f o r the moment), the main p o i n t o f c on tr a st
between the two approaches w i l l be, in pra ctic e , the issue o f fuzzy
boundaries. Otherwise, componential a na ly s is often works with
prototypical c ondi ti ons de s p i t e the theoretical adherence to the
'maximal generalization' ideal, i.e. to the requi rement that the

- 39 -
d e f i n i t i o n must be general enough to cover a l l pos si b le instances o f
the item.

To f u l l y a ppr ec i at e the n egat i ve r o l e this adherence has played in

the c o ns t ru ct io n o f d e f i n i t i o n s , the confusion concerning l i n g u i s t i c
and e x t r a ! i n g u i s t i c i nf or ma ti on and the r elevance o f these problems
to the d e s c r i p t i o n o f motion verbs, a second example w i l l be looked
at in some d e t a i l , focusing on c h e c k l i s t definitions for 'run' and
some r e l a t e d motion verbs.

M i l l e r and Johnson-Laird (1976) gi ve a general formula f o r ' wa l k ' of

the form: ACT (x,S') and CAUSE (S', DO ( FEET, S)) and ALLOW
( S, TRAVEL(x)) . Even i f you add (ON(TRAVEL)) (x,LAND) you s t i l l have
at l e a s t 'run' and a l l its hyponyms plus a l l the hyponyms o f ' w a l k '
satisfying these c o n d i t i o n s . So t h i s is a general formula for all
'travel-on-foot Vs', as they c a l l them, although it actually first
appears as a tentative definition for 'w alk'. New, M iller and
Johnson-Laird clai m t h a t the basic d i s t i n c t i o n w i t h i n a l l 'travel-on-
foot' Vs is between 'walking' and 'running'. So they consider
incorporating a component 'RAPIDLY' in 'run' which, however, they
r e j e c t on the grounds t h a t sentences such as He walked r a p i d l y and I
was forced t o run slowl y to keep up wi th him are not s em an ti c al l y
anomalous. Hence their final proposal i ncludes an oper at or ATG
(FEET) f o r 'walk' (ATG=always touching ground); 'run' gets ' n ot ATG'
and a l l is s e ttle d . We can r e co ns t r u c t a d e f i n i t i o n f o r ' r u n ' on the
basis they provide and get:

(WITH,(not ATG(ACT)) ) (x,S,FEET) and CAUSE (S,(ON(TRAVEL))

(x, LAND).

The f i n a l component is considered omi ssi bl e from the corresponding

definition for 'walk' because we can understand ' wa lk on a i r ' : "the
conventional restriction to land might be considered part of our
general knowledge rather than our l i n g u i s t i c knowledge" ( p . 552) we
are told. What we are not t o l d is that 'fe e t' is equally a
'convent ional restriction' and arguably also part of our ' general
knowledge'; and, besides, that a person w it h a r t i f i c i a l limbs or on

- 40 -
crutches is most certainly 'w a lk in g '. What is worse, if we are
really l ooking for the most general (and 'unconventional')
definition, then TRAVEL i s also omi ssi bl e, since one can ' r u n ' on the
spot, without advancing at a ll, for physical exercise - unless, of
course, we would again like to have two e n t r i e s 'run^' and 'run2 '
(where t he second, incidentally, and not the firs t one, w ill be
s p e c i f i e d as l a c k i n g the TRAVEL component).

S i m i l a r l y Wi erzbi cka (1972) c r i t i c i z e s Wotjak (1971) and Baumgartner

(1967) for treating 'walk' and ' r u n ' as i d e n t i c a l except f o r 'speed'
n o t i c i n g t h a t you can walk f a s t and run slo wl y. Her argument is "The
difference in the re lative speed between wal ki ng and running is a
consequence o f the q u a l i t a t i v e difference between t h e two types of
movement, and t h e r e f o r e need not be marked at a l l , provided t h a t the
q u a lita tiv e difference its e lf is captured" ( p . 1 0 7 ). But Wotjak has
to fight agai ns t structuralist redundancy and stick to one-word
features (as best he can) at the same t ime . He is working wi th
semantic criteria - semantische Kriterien (Merkmale) - such as
'a nim ate', 'agent', ' dy na mi c' , etc., and might be t r y i n g to avoid
c o l l a p s i n g ' l a u f e n ' wi th ' rennen' and both wi th ' s p a z i e r e n ' and t h e i r
respective hyponyms. Conditions such as 'in term itten t' vs
'c on ti n uo us ' c ont act or ATG may not constitute c o nt r a s t s useful
throughout the paradigm, w h i le ' v e l o c i t y ' helps wi th hyponyms and, in
his view at l e a s t , also w i t h near synonyms ( ' l a u f e n ' gets + / - fast,
w hi le 're nne n' gets + f a s t ) . What is even worse, unless t he motion
i s s p e c i f i e d as continuous, t h e r e seems to me to be no way o f keeping
' r u n ' separate from 'jump' or 'hop' even.

Far from trying to ridicule serious attempts at satisfying the

checklist ideals, I am onl y attempti ng to show the consequences o f
t he adherence to these principles in one of the best cases for
s t r u c t u r a l i s t analysis: 'manner' s pec if yi ng motion Vs. If one wants
to impose structure (try to establish more complete p a tt e r n s than
t h e r e i s evidence f o r ) , one can do i t . I t is, however, u n l i k e l y t h a t
the r e s u l t w i l l be worth the e f f o r t .

- 41 -
 t

If definitions are to be interesting and illum inating they must

somehow account f o r speakers' intuitions. Semantic competence i s not
discovered by simply saying t h a t He was running slowl y and I had to
walk f a s t ... ( o r the o t he r way round) is not anomalous. One does
not need 'bizarre c o n te x t s ' to r e a l i z e that velocity is a relative
matter. And as for the sentence its e lf, it only indicates that
someone can be wal ki ng f a s t e r than usual t o keep up w i t h someone who
is running more slowl y than usual. If anyt hi ng, this particular
sentence i mpl ie s that the 'standard' instances of 'walk' are
certainly slower than the standard instances o f 'ru n'. Juxtapose a
corresponding sentence w it ho ut adverbs, e . g . He was running and I had
to walk t o keep up wi th him and one sees immediately t h a t i t re qui re s
consi der able computing to understand lite ra lly, precisely owing to
the difference between 'walk' and 'run' in speed. And we do not
really need speci al sentences to t e l l us that. Under normal
' ev er yda y' circumstances people run i f they want to go (on foot)
somewhere f a s t . Otherwise they walk.

Evidently, all this is e x t r a ! i n g u i s t i c i n for ma ti on but i t may happen

to be important i n for ma ti on in distinguishing between 'run' and
'w alk'. As a lr e ad y mentioned, no pr ope rty that is necessary to
keep these verbs a pa rt i s n e c e s s a r i l y ' p u r e l y ' linguistic. I t could
o f course be argued t h a t 'speed' in ' r u n ' is a concomitant f a c t o r o f
the type of motion i t d e sc ri be s. But the argument could be reversed.
What i s t he re to d i c t a t e the order? It could also be the case t h a t
'manner' i s a concomitant f a c t o r o f the i n t e n t i o n f o r f a s t movement
in combination wi th the c a p a b i l i t i e s o f the human body. Notice also
t h a t 'manner' i s not as r e l a t i v e to p a r t i c u l a r i n d i v i d u a l s as 'speed'
is. In t h a t respect it may be considered, more 'b a s ic ' or simpl er .
Be t h a t as i t may, MG ' t r e x o ' (run), at le a s t, arguabl y involves a
speed component in its commonest understanding; it is figuratively
used for any kind of fast a ctivity: talking fast, driving fast,
working fast, and I take this to be 'p u re ly' linguistic evidence.
Yet 'tre x i (p o li) vr/vora' ( s / h e runs (v er y) fast) i s not redundant
nor i s the exact t r a n s l a t i o n o f M i l l e r and Johnson- La ir d' s sentence
( He was walking f a s t . . . ) anomalous.

- 42 -
N ot i c e , in this connection, that M iller (1972) also considers
i n c lu di n g v e l o c i t y in the components o f ' r u n ' and expresses doubts,
on the basis o f examples such as the exi ste nce o f ' j o g ' , which equals
'run s l o w l y ' . It i s q u i t e possi bl e t h a t f o r an adequate d e s c r i p t i o n
the taxonomic levels need to be e s t ab li s he d firs t and whatever
s t r u c t u r i n g t h er e e x i s t s w i t h i n each one s e p a r a t e l y di scover ed. For
since ' v e l o c i t y ' is a r e l a t i v e matter, 'jog' i s not t o be contr ast ed
to 'w alk', for i ns t an ce , but t o the ot he r hyponyms o f 'ru n '. Once
relatively 'fa s t' motion is established for 'ru n ', 'jog' is to be
understood as implying ' l e s s fast' in connection t o t he ' h i g h speed'
of ' r u n ' .

These f a c t s can be most n a t u r a l l y expl ai ned only by r e f e r e n c e to the

prototypical understanding of 'run' (or, at least of 'tre x o ' for
which I have more r e l i a b l e e vi d enc e) . Only t h a t kind o f d e s c r i p t i o n
can contai n ' f a s t movement' ( or even the i n t e n t i o n f o r f a s t movement)
as an appropriateness c o nd i ti o n and accommodate the fact that 'run
s lo wl y' is not anomalous, link it wi th some prototypical
understanding o f human motion and wi th the absence o f a s p e c i f i c a t i o n
of 'speed' for 'w alk'. Other elements such as ' c o n t i n u i t y / d u r a t i o n '
are also prototypical c o ndi ti ons for 'walk' and 'ru n ', as
'instantaneous' is f o r 'jump'. They are probably also necessary f o r
a most general -mi ni mal definition i f the verbs in question are to be
kept separ ate; the problem of redundancy does not arise: She was
walking conti nuousl y does not seem to be s u f f e r i n g from redundancy.

Notice also t h a t n o n - p r o t o t y p i c a l t reatments o f 'jump' include OVER

in the definition; M iller and Johnson-Laird explain i nstances of
'jump under' as p a r t i c u l a r cases o f a general JUMP OVER. They also
accept t h a t i nanimate obj ect s can 'jump' so t h a t ' i n t e n t i o n a l i t y ' and
'instrum entality' ( i nc luded in their definition) could be omi tted.
So why do these f e a t u r e s appear in the general formula in the f i r s t
place? Because t h e i r definition is supposed t o correspond to "the
sense t h a t children learn first" ( p . 5 5 7 ). Such i n for ma ti on is very
important and cannot be overlooked. I t might w el l be the case t h a t
the ' un a bb r ev ia t ed ' d e f i n i t i o n i s also close to the most p r o t o t y p i c a l
understanding o f ' j u m p ' . For c e r t a i n reasons it corresponds to the
most 'natural' and straightforward instances of 'jump'. Similarly

- 43 -
f o r ' l e g s ' or ' f e e t ' in connection with ' w a l k ' and ' r u n ' . Wierzbicka
expresses reservations on the grounds that "one can walk on one's
hands" (1972:108). And she is right. Wi thi n the system she is
a p pl yi ng, 'most g e n e r a l ' has nothing to do wi th 'most n a t u r a l ' . Even
the standard paraphrase o f 'w alk', i.e. 'g o-on-foot', is not v a l i d .
Alternatively walk on one's hands w i l l be t r e a t e d as a metaphor,
which i t i s not e x a c t l y . Now, if 'legs' (or 'fe e t') disappears from
the d e f i n i t i o n of 'w alk', then 'wa lk on f o o t ' should not r e a l l y be
redundant (which i t is).

For a 'prototypical' tr e at me nt of 'walk' these facts would not

constitute a problem. I n cl us i on of legs/feet seems to be an
obligatory c o n di t i o n for 'w alk'; hence the redundancy of 'walk on
fo o t'. On the o th e r hand, walk on one's hands w i l l be understood as
a d e v i a t i o n from the pr ot ot ype : n e i t h e r anomalous, nor f i g u r a t i v e .

It seems to me t h a t Fillmore's (1975:129) comment: "introspection

about a p p r o p r i at e language use in bizarre contexts does not y i e l d
highly dependable data for semantic research" becomes more
interesting if one considers the sort of data-base used for
e s t a b l i s h i n g most general -mi ni mal d e f i n i t i o n s . For I cannot t h i n k o f
a goal t h a t would f or ce the s ema nt ic ist to work on t he basis of more
atypical, unusual and ' b i z a r r e ' conte xt s.

I have so f a r limited the comparison o f what seem t o me the most

important approaches t o l e x i c a l semantics to 'p ra c tic a l' issues, in
an e f f o r t to show t h a t a number o f important l e x i c a l semantic f a c t s
could, in principle, at least, be accounted for within a
'prototypical' framework, e s p e c i a l l y since p r o t o t y p i c a l c ondi ti ons o f
a p p l i c a t i o n need not be s t a t e d in e l l i p t i c and n o n - d e s c r i p t i v e terms.
Limiting the semantic specification of lexemes to metalinguistic
terms and 'neat' formulae l e ads , in a number of domains, to
uni n fo r ma t iv e s e m i - d e f i n i t i o n s .

- 44 -
1.2.6 S t r u c t u r a l i s m vs Prototype t heor y: c o n c e p t u a l i z a t i o n and
extra!inguistic re a lity

It has a lr e ad y been pointed out t h a t structuralism and Prototype

theor y ta ke a different stand on the issue of conceptualization
concerning word meanings and human c a t e g o r i z a t i o n in p a r t i c u l a r . The
cornerstone of structuralist semantic theories is probably the
interdependence of word meanings and t h e r e f o r e semantic cont ra st s are
t he focus o f t h e i r a t t e n t i o n .

Evidence from language a c q u i s i t i o n m a y .: reinforce this view,

(Lyons 1983:63), Contrary to this, there are arguments for the
independence of word meanings favoured by some proponents of
Prototype t he or y (e .g. Verschueren 1981), as a lr ea dy mentioned. As
evidence it is claimed t h a t people, when asked to d e f i n e words, do
not f e e l obl iged to r e f e r to r e l a t e d items, but u s u a l l y concentrate
on the f ocal characteristics o f the item in quest ion, i.e. descri be
the p r ot ot y pe . Accor dingly, only genuine structuralist definitions
are based on contrasts. Yet any theor y should be able to
differentiate between words which are not synonymous f o r the sake of
elementary d e s c r i p t i v e adequacy. Theref ore the p o s s i b i l i t y t h a t f o r
some words a t l e a s t and a t some l e v e l at l e a s t the d i s t i n c t i o n s are
contrastive, cannot be r ul ed o u t. Besides, to do Prototype theory
justice, one must emphasize t h a t words are not understood as i s o l a t e d
entities in the abstract. They are understood as being r e l a t e d to
o t h e r words and, quite importantly, as r e l a t e d t o the world (extra-
linguistic reality). S che ma ti ca ll y one could say that w hi le
s t r u c t u r a l i s m i s i n t e r e s t e d in the former ( i n t r a l i n g u i s t i c ) r e l a t i o n ,
Prototype t he or y i s i n t e r e s t e d in br i ngi ng out t he l a t t e r one through
the recognition of the cognitive pa tt e rn s of categorization. The
main premise in this r espect has al re ady been mentioned: word
meanings are t o some e x t e n t i n d e te rm in a te ; communication is achieved
through c on ce n tr a ti o n on the focal points, i.e . the most
characteristic (prototypical) instances o f a cat egory.

- 45 -
It is very interesting to consider the structu ralist stand on the
issue o f indetermi nacy. Its c l e a r e s t e xp os i ti o n is probabl y Coseriu
and Geckeler (1 98 1: 4 9, c i t i n g Coseriu 1966):

" l es val eur s 1 i n g ui s t i q u e s sont des val eur s conce ptuel l es qui se
d 6 f i n i s s e n t par l e u r s opposi ti ons e t par l e u r fonctionneinent, e t
non pas par des c r i t e r e s ' r e e l s ' e t par l es l i m i t e s , pr eci ses ou
i mpreci ses, e n t r e l es phenomenes de l a r e a l i t e " .

So, d ifficu lties i n the separ ati on o f r e al phenomena are explicity

said not to affect the distinction between the corresponding
c oncepts:

"quite the contrary: such difficu lties show t h a t the concepts

are clearly separated. Thus, e.g . the fact that in
extra!inguistic reality there are no clear boundaries between
day and n i g ht does not mean t h a t the concepts 'day' and ' n i g h t '
are unc lea r as concepts. Here, th erefore, the precise
delim itation of the concepts stands in opposi ti on to an
imprecise delim itation of the phenomena condi ti oned by the
nature o f e x t r a ! i n g u i s t i c d a t a " .

(Coseriu and Geckeler 1981:49)

Hence it is claimed that language e s t a bl i s h e s boundaries in areas

which e x i s t as a continuum and the s p e c i f i c example o f f e r e d i s colour
adjectives. But B e r l i n and Kay (1969) and Kay and MacDaniel (1978)
have shown that there is a correspondence between the internal
structure of r e a l i t y and the laws determining t he c ol our s e n s i t i v i t y
of the eye and that the linguistic choice (the p oi nt s of
lexicalization) corresponds to the focal poi nts o f t he f uz z y (as to
their boundaries) areas o f the col our spectrum. Hence the apparent
arbitrariness of c ol our terminology is only a ppar ent. Now i f the
linguistic choice is not completely arbitrary but its e xpl ana ti on
rests with e x t r a l i n g u i s t i c information, then t h i s i n f or m at i on cannot
be ignored and word meanings cannot have r e p r e s e n t a t i o n s which are a t
odds wi th the r e a l world phenomena t h a t would e x p l a i n them.

- 46 -
Consider now Labov (1978:220) on the issue o f indeterminacy:

"I n the most general sense, l i n g u i s t i c anal ys is i s the study o f

l i n g u i s t i c categories. The l a r g e s t p a r t o f our e f f o r t i s devoted
t o di sco ve ri ng c a t e g o r i e s , d e f i n i n g them and s e t t i n g up r u l es f o r
assigning membership in those c a t e g o r i e s . The e n t i r e a c t i v i t y is
dependent on the exi ste nce o f category boundaries i f i t i s to be
meaningful: i f t h er e are no e f f e c t i v e boundaries between two
c a t e g o r i e s , the assignment o f members to one or another is
obvi ousl y an a r b i t r a r y and p o i n t l e s s e x e r c i s e " .

Labov distinguishes between domains in which features operate

categorically and others in which only probabilistic or weighted
features are operative. His examples include, on the one hand,
k inshi p terms where he expects c a t e g o r i c a l judgments "since they are
ascribed statuses t h a t do not change over time" ( i b i d . :226) and, on
the o t h e r , ' achieved s t a t u s es ' like 'adult' where weighted judgments
are r e q u i r e d . It seems to me t h a t even w i t h i n the same domain the re
are poi nts lexicalized along a single axi s (which is probably the
most obvious and simple ' br e ak in g up' of a continuum) and points
scattered in different directions but related to one another v ia
certain dimensions. Some of these relations may be contrastive
almost in the s t r u c t u r a l i s t d i s t i n c t i v e f e a t u r es sense. Phenomena of
gr adat io n and indeterminacy can be expl ai ned on the basis of
prototypical/stereotypical approaches. Structuralist methods can,
however, also be useful in establishing a basic pattern, without
adhering to the view t h a t the value o f a word i s determined by the
s t r u c t u r e o f the whole f i e l d .

In very concrete terms consider the examples of motion verbs

p r e v i o u s l y discussed. If some event is not e x a c t l y an i nstance of
'tre x o ' (run), i.e . not prototypically a 'trexo' event but almost
t h a t , the d e v i a t i o n w i l l not be completely u n p r e d i c t a b l e , i.e. i n any
direction. In the vast majority of cases it is quite predictably
going to be in the direction of the nearest categories, i.e. in
specific directions: either a 'p e rp a ta o ' (walk) event - but not
e x a c t l y t h a t - or a ' p i b a o ' (jump) one - but not e x a c t l y t h a t . Other
p o s s i b i l i t i e s might be 'hop' or 'dance' (and then the actual di stance
from the protot ype might be g r e a t e r ) but c e r t a i n l y not ' s i t ' . Focal
poi nts may be established on the basis of the prototypical

- 47 -
characteristics of a category and at the same t ime a s t r u ct u r e d
r e l a t i o n s h i p may be de tect ed between r e l a t e d focal p o i n t s .

So the prototypical descriptions of 'tre x o ', 'p e rp a ta o ' , 'p ib a o '

would probably have to i ncl ude as main poi nts of contrast the
condi ti ons detect ed but sometimes misused by s t r u c t u r a l i s t methods
(e .g, continuity o f motion, number o f feet on the ground at given
moments, etc.). They would predict that for marginal instances
subj ects hesitate to use 'tre x o ' (r un) and f i n a l l y opt f o r 'p ib a o '

(jump) or 'p e rp a ta o ' (walk) depending on which o f these c ondi ti ons

i s / a r e absent, but do not use a f o ur t h completely u nr e l a t e d element.

An example o f what I c a l l the s impl est breaking up o f a continuum i s

o f f e r e d in the domain o f MGMVs by ' k s a p l o n o 2 r ( l i e down) - '/faSome^'
(sit down) - 'sik o n o m e ' (stand up, rise). Some o f the in between
stages are also l e x i c a l i z e d : '\/e rn o ' (lean), ' misokaBome' (h a lf-s it),
' m isoksaplono' (h alf-lie ), ' m isosikonom e' (half-stand). All these
are poi nts on a single axi s ( po in t s of body touching supporting
surface), so no f u r t h e r s t r u c t u r i n g needs to be evoked; the d e v i a t i o n
from the focal poi nts is in a single direction, the pattern is
u ni d im ens iona l.

The above c o ns ide ra ti ons bear on the s o rt o f apparatus one needs f o r

a particu lar description. A prototypical a n al y si s o f Vs l i k e 'buy'
and ' s e l l ' might r e q u i r e r ef e re n ce to scenes, s ce nar ios, frames and a
number of sim ilar tools used by prototypical semantics, but quite
different c o nsi de ra ti ons would be useful for motion verbs. The
prototypical scene f o r ' kaBome' (s it) might i nclude a c h a i r and f o r
'k s a p lo n o ' (lie ) a bed, but the role these elements p l ay in the
a p p l i c a t i o n o f the terms in question may be l e ss c r u c i a l than t h a t o f
a c o ndi ti on such as 'knees bend' proposed by Dahlgren (1978) f o r the
s te reot ype o f ' c h a i r ' and e q u a l l y useful f o r the ste reot ype o f ' s i t ' .

it is also r a t h e r c l e a r t h a t by t h e i r very nature c e r t a i n f e at ur es

are graded, i.e. they are a m at te r o f m o r e - o r - l e s s , w h i l e others are
not , i.e. they are a ma tt er o f yes-no. Consequently it is quite
possi bl e t h a t f o r the d e s c r i p t i o n o f the semantic f a c t s o f a s i n g le

- 48 -
lexical item one needs to have recourse t o both kinds of c r i t e r i a (as
w ill be shown t o be the case wi th MGMVs).

F i n a l l y j u s t as one does not want a theory to a r b i t r a r i l y and r i g i d l y

impose schemata which do not r e f l e c t r e a l i t y , one s i m i l a r l y does not
want a model which allows for anything at risk of becoming
super fluous. Prototype theor y i s discussed in the f o l l o w i n g secti on
wi th a view to clarifying which of its tenets are useful for a
d e s c r i p t i o n o f MGMVs. The o v e r a l l picture, f o r the moment, seems to
be t h i s : s t r u c t u r a l i s t theories risk imposing s t r u c t u r e and l e av i ng
out phenomena o f gr adat io n and i ndeterminacy; Prototype t heory r i s k s
not accounting for existing structure by prov idi ng open-ended
definitions conflating semantic, associative and individualistic
information. Naturally, whether these appar entl y d i f f e r e n t kinds o f
inf or ma ti on can be kept separate i s a matter o f debate.

1.3 Prototype t he or y and human c a t e g o r i s a t i o n

1.3.1 On Putnam's stereot ypes

In the previous s ec ti on it was mentioned that Prototype theor y

proposed identifying the prototypical i nstance of a category and
a ll owi ng f o r o the r instances to be members o f t h a t category wi thout
exhibiting all the identifying criteria . This t heor y has been
juxtaposed to theories of word meaning which do not a ll o w for a
specific item to be a legitim ate i nstance of a c at egor y unless a
number o f necessary and s u f f i c i e n t c o n d i t i o n s / c r i t e r i a are f u l f i l l e d .
It has not y e t been pointed out here t h a t the basis f o r t h i s r a d i c a l
difference o f p o i n t o f view i s , in e f f e c t , the contr over sy between
e xt ensi onali sm and i n t e n s i o n a l i s m . ^

In t he present c ont ext only a couple o f poi nts need t o be discussed

at some l e n g t h , starting with the issue o f a n alyticity. Putnam's
(1970, 1975) t he or y of stereot ypes w ill be j uxtaposed to those o f
Quine and Katz wi th respect to this problem. The most convenient

- 49 -
s hor tc ut in doing this seems t o be a c o n si d er a ti o n of the age- ol d
examples o f the type:

(a) cats are animals

(b) bachelors are unmarried.

Roughly speaking, while for Quine n e i t h e r (a) nor (b) is analytic,

for Katz they both a r e. Putnam would clai m t h a t type ( a) examples
are not a n a l y t i c on the assumption t h a t the pr o pe r ty o f "animalhood'
i s r e v i s a b l e and t h a t in case i t i s a c t u a l l y r evi sed (i.e . absent or
replaced by "mechanical obj ect" for i ns tance ) the meaning o f "cat"
does not change; what does change i f i t turns out t h a t " c at s ' are in
fact mechanical objects, for i n st an ce , is our beliefs about them.
K a t z ' s stand on t h i s po i nt i s c l e a r l y t h a t in such a case e i t h e r the
meaning o f "cat" changes, or we s t a r t using a new name f o r t h i s kind
o f "creature" (Katz 1 97 5 :9 8) .

It i s f a i r l y easy t o demonstrate t h a t Katz i s wrong about t h i s l a t t e r

c l ai m , but hi s former one presents a much more seri ous problem. To
put i t very simpl y, it seems q u i t e u n l i k e l y t h a t t he day we discover
that "cats' are a c t u a l l y mechanical obj ect s we w i l l coin a new name
for them; once the r e f e re n ce is fixed (and named) the name (term)
continues to be attached to the object. Pulman (1983) offers the
example o f "whales" which were not renamed a f t e r it turned out t h a t
they were not " f i s h ' but 'mammals'. Perhaps t h i s p a r t i c u l a r example
is not convincing enough because the d i st an ce between 'fis h ' and
'mammal' i s not as g r e a t as the one between ' a n i m a l ' and "mechanical
object' (c ons ider the common sense difference between hi gher and
l ower level features), or because "whales' are probably among the
poorest examples of 'mammals' in the specific prototypical/
stereotypical sense a l re a dy mentioned and discussed in more d e t a i l in
the present section. Consider, however, the example o f "spaghetti'
which was a p p ar e n t l y 'discovered" one day ( 1 s t o f A p r i l ) to be not an
"industrial product" but a . . . "plant". It is quite unlikely that
the meaning o f "spag het t i" changed f o r those speakers ( A p r i l ' s f o o l s )
who were "informed" about its actual n a tu r e , and q u i t e impossible

- 50 -
t h a t they would have s t a r t e d t h i n k i n g about a new name f o r i t . What
must have changed, f o r a w h i l e , was t h e i r b e l i e f s about i t .

This links in an interesting way to Putnam's distinction between

'su p e rficia l' and ' e s s e n t i a l ' p r o p e r t i e s and hi s c lai m t h a t knowledge
o f the former r a t h e r than the latter is crucial t o determini ng the
meaning of an entity for the average s p e a k e r . I take this
d i s t i n c t i o n t o be an i mportant c o n t r i b u t i o n o f Putnam's t h e o r i z i n g t o
a b e t t e r understanding o f word meaning and t h e r e f o r e worth looking at
rather closely. Putnam concentrates on n a tu ra l kind terms and claims
that their superficial properties are central to their meaning.
These are juxtaposed to essential properties, i n c lu s i o n of which
would guarantee membership in t he extension o f a term, but which are
not d i r e c t l y l i n k e d to an i n d i v i d u a l speaker' s competence.

From a c e r t a i n p oi n t o f view t h i s may be understood as saying simply

that we have to abandon the search for necessary and sufficient
conditions. In fact, Putnam's ideas in this r espe ct are
diametrically opposed to those of Katz and Fodor; the latter are
commonly understood as clai mi ng that necessary and sufficient
condi ti ons f o r membership in t he extension o f a term such as those
Putnam considers ( i . e . 'gold', 'aluminium', 'w ater') are ( i m p l i c i t l y )
known by any speaker who uses the terms a p p r o p r i a t e l y : it should be
r e c a l l e d t h a t both semantic markers and d i s t i n g u i s h e s were supposed
to r epr esent exhaustively a speaker' s competence f o r the r e sp e c t i v e
term. Putnam, on the o t h e r hand, claims t h a t a speaker' s competence
which enables hi m/ her to use a term appropriately cannot involve
necessary and s u f f i c i e n t c ondi ti ons f o r membership i n the ext ensi on,
because these are not known to a ll, or even any speakers.^ He
actually suggests t h a t the individual speaker' s competence i ncludes
e ve ry thi ng except the e xt ens ion. A speaker can be sai d t o know the
meaning of 'ca t' on the single c o nd i ti o n that s/he knows the
superficial or standard, i.e. 'stereo typ ica l', characteristics of a
cat. S/he can be said to use the term a p p r o p r i a t e l y if h i s / h e r use
"passes muster" and the extensi on of 'cat' in his/her idiolect is
actually the s et of cats. The extension is not fixed by what the
individual speaker grasps or not , but by the community, i n c lu di n g the
e x pe r ts , through a complex c o - o p e r a t i v e process.

- 51 -
So Putnam c r i t i c i z e s ot he r t h e o r i e s which, in hi s words, "l eave out
the c o n t r i b u t i o n o f s o c i e t y and the c o n t r i b u t i o n o f the r e a l world"
( p . 1 64), i.e. understand c o gni ti on as a p u re l y individual m a tt e r,
ignor ing its s oci al dimension. Contrary to t r a d i t i o n a l philosophy,
extension is t o be determined s o c i a l l y (contribution of society) and
in part indexically ( c o n t r i b u t i o n o f the w o r l d ) . It depends on the
actual nature o f p a r t i c u l a r e n t i t i e s , and t h i s actual "hidden" nature
is not fu lly known to the speaker. This does not imply that
extension i s not a component o f meaning; i f in using the same term we
refer to entities with different extensions we actually mean
d i f f e r e n t things. Ne ve rt he l es s, t h i s d i f f e r e n c e in extension is not
a r e f l e c t i o n o f a d i f f e r e n c e in i n d i v i d u a l psychological states, i.e.
accountable f o r in terms o f l i n g u i s t i c competence.

Putnam's proposal f o r the semantic r e p r e s e n t a t i o n o f a na tur al kind

term involves a lis t of 'syntactic ma rke rs' , 'semantic ma rker s' ,
'stereotypes' (i.e. sterotypical properties) and ' e x t e n s i o n ' . These
constitute a hypothesis about the individual speaker' s competence
w it h the notabl e exception o f the 'extension'. This representation
he c a l l s the 'normal form d e s c r i p t i o n ' of a term and the specific
example o f such a d e s c r i p t i o n he o f f e r s f o r ' w a t e r ' looks l i k e t h i s :

A. s y n t a c t i c markers (box)
mass noun, concrete
B. semantic markers (box)
n at ur al ki nd, l i q u i d
C. s te reo typ e /'s te re o ty p ic al properties' (box)
colourless, transparent, ta steless, t h i r s t quenching
D. extensi on (box)
H2 O ( g i v e or take i m p u r i t i e s ) .
(Putnam 1975:269)

This r e f l e c t s quite c l e a r l y the d i s t i n c t i o n between s u p e r f i c i a l and

essential properties. Superficial properties, which are included in
the s t e r eo t y pe , are p a r t o f the common s p e a k e r - h e a r e r ' s understanding
of the term and they can be "wild ly i na cc ur ate" or even
"scientifically wrong", since in Putnam's terms, a s te reot ype is a

- 52 -
"conventional (frequently m al ic ious) idea (which may be w ildly
i n a cc u r at e ) of what an x looks like or acts like or is ..."
( i b i d . :249).

Essenti al properties, which constitute necessary and sufficient

c ondi ti ons f o r membership in the ext ensi on, are known to s p e c i a l i s t s ,
or e x pe r ts , but are said to be all present in the linguistic
community considered as a collective body; that collective body
d i v i d e s the ' l a b o u r ' o f knowing and using the pa rts o f the meaning o f
a term, through a s o c i o l i n g u i s t i c process Putnam c a l l s "division of
linguistic l abor" ( i b i d . :144), w ithout s p ec i f y i ng how exactly it
operates (leaving the ma tter to sociolinguists apparently). The
whole hypothesis i s o f f e r e d as an expl anati on o f the f a c t t h a t whi le
only a minority of speakers has special (scien tific or ot he r)
knowledge concerning certain terms, all or most speakers of a
linguistic community understand and use these terms more or less
appropriately.

Putnam's contribution can perhaps be fully appr eci ated if one

concentrates on hi s own examples: 'gold', ' a l u m in i um ' , 'w ater'. In
such cases the increase o f s c i e n t i f i c knowledge in r e l a t i v e l y recent
years has separated e xpe rt from non-expert understanding o f the terms
rather dramatically, wi thout these words having n e c e s s a r i l y changed
meaning in terms o f ' i n d i v i d u a l l i n g u i s t i c competence'; i t is hard to
imagine t h a t 'w ater' changed meaning f o r the average speaker- hearer
once it was discovered that it was actually HgO. Improvement of
techniques brings s o c i e t y as a whole c l o se r to a s a f e r i d e n t i f i c a t i o n
o f the e s s e n t i a l properties, the "hidden s t r u c t u r e " ; y e t the average
s pea ker -hea re r re lie s , at any given ti me, on the 'currently
ope r at io n al definition' of a term. Two r e l a t e d questions arise at
th is point:

1. How to determine the minimum l e v e l o f competence (so t h a t i t

covers t he currently ope rat io nal definition of a term) vs
i nf or ma ti on in the extension box.

- 53 -
 2. What happens wi th common terms which do not have a
s c i e n t i f i c a l l y / t e c h n o l o g i c a l l y v e r i f i a b l e e xt e ns i on.

The common basis o f these questions/problems is probably t h a t Putnam

does not provide anything l i k e a d e t a i l e d t he or y o f stereot ypes (nor
does anyone e ls e y e t , as f a r as I know). His general view is t h a t
t h i s 'minimum l e v e l o f competence' depends on the c u l t u r e , on the one
hand, and the natur e o f the object, on the other, but t h a t w i t h i n
these vast lim its any n a t i v e speaker i n t u i t i v e l y knows what belongs
to it. Hence what is central to the meaning of a term is this
s t e r e o t y p i c a l , w i de ly known, sometimes i n c o n s i s t e n t and e x t e n s i o n a l l y
in correct, but l i n g u i s t i c a l l y obligatory i n f o r m a t i o n . So a l l that
can be detect ed in hi s relevant te xts in this connection is two
partial but very important answers:

( a) The ste reot ype need not be c o r r e c t or unchangeable (i.e .

un re vi sa bl e or a n a l y t i c ) : "linguistic obligatoriness i s not
supposed t o be an index o f u n r e v i s a b i l i t y or even o f t r u t h .
So, for i n st an ce , we can hold t h a t 'tigers are s t r i p e d ' is
p a r t o f the meaning o f ' t i g e r ' wi thout being trapped in the
problems o f a n a l y t i c i t y " ( i b i d . :177), i . e . w i t ho ut having to
accept t h a t ' t i g e r s are s t r i p e d ' is a n a ly tic .

(b) For many words an e x t e n s i o n a l l y c o r r e c t t r u t h d e f i n i t i o n is

in no sense a t heor y o f the meaning o f the word. Hence,
discussing Davidson's t h e o r y , Putnam c e r t i f i e s t h a t " ' w a t e r '
is true of x iff x is HgO" ( i b i d . : 1 80 ), although
extensionally correct, would tell us nothing about the
meaning of 'water' if most speakers did not know that
'water' i s H2 O.

Linguistic e xp l an a ti o n is at least di sent angl ed from the

phi losop hi ca l problems o f a n a l y t i c i t y . I f the semantics o f a term is
to be a r e p r e s e n t a t i o n of the average speaker' s knowledge of this
term, it makes much more sense to concentr ate on the noti on of
lin g u istic obligatoriness (i.e . on s p e c i f y i n g the 'minimum l e v e l of
competence') .

- 54 -
Now the i d e a l i z a t i o n i m p l i c i t in the noti on o f the ' average speaker'
i s a problem in i t s e l f as w i l l be demonstrated in the course o f t h i s
i n v e s t i g a t i o n time and time a gai n. I t seems t h a t f o r d i f f e r e n t kinds
of common terms, their meaning w ill d iffer depending not on a
'lin g u istic community' i n the broad u n q u a l i f i e d sense in which the
term i s f r e q u e n t l y used, but on what s o r t o f i n fo r m a t i o n i s a v a i l a b l e
at any given time to a much more restricted subgroup of a given
linguistic community. These subgroups might have to be de fi ne d on
the basis of their educati onal level, specific pr of e ss iona l or
s c i e n t i f i c knowledge, l o c a l characteristics, e tc .^ Although t h i s is
mainly a s o c i o l i n g u i s t i c problem, i t i s o f immediate concern f o r t h i s
kind of semantic t h e or y . This i s c l o s e l y linked to the issue of
communication regar di ng word meaning. It seems h i g h l y d e s i r a b l e to
be in a p o s i t i o n to t e l l a t which points communication breaks down
and f o r which reasons it does. This can be the object of further
research in a number o f interrelated disciplines. It seems to me,
however, that Putnam's t he o ry , incomplete as it is, poi nts to the
right direction of where to look for answers. Stereotypical
information, which i s respons ibl e f o r communication, depends on the
nature of the object for which a term is used. I t is therefore
p r e d i c t a b l e t h a t f o r kinds o f terms ot he r than those Putnam analyses,
the sociol i n g u i s t i c division of labour may not e x i s t at a ll: t h e re
may be no experts who can decide in case of doubt, or they may have
d i f f e r e n t opi ni ons. A good case i n po i nt i s a term l i k e 'democracy'
which involves different s oci al theories. It is predictable that
notions related to such terms, far from being accountable for in
terms of necessary and sufficient conditions, w ill constitute
p r e c i s e l y poi nts a t which communication may wel l break down.

It must be noted, however, that asserting that stereotypical

i nf or ma ti on depends on the nature o f the o bj e c t f o r which a term is
used i s only a s t a r t i n g point. It i s r a t h e r e vi d e n t that relative
salience of external (superficial) characteristics has to be taken
seriously into consideration. So, even within the same semantic
field , a single s peaker 's linguistic competence w i l l be d i f f e r e n t l y
de fi ne d concerning 'beech' and 'elm trees', on the one hand, and
'palm t r e e s ' , on the o t h e r . Nature o f the o b j e c t i s probably to be
understood much more narrowly than Putnam seems to suggest. In

- 55 -
a dd i t i o n to t he natur e o f the o b j e c t , Putnam also recognizes t h a t the
r o l e an o bj e c t plays in a p a r t i c u l a r s oc i e t y w i l l a f f e c t the type and
amount o f s t e r e o t y p i c a l information. Yet even these parameters are
not sufficient. Compare 'elm trees' to 'tig e rs ' for i ns tan ce.
N e i t h e r i s probably o f g r e a t importance to Western urban popul ati ons ;
in fact the latte r category is remoter than the former one. Yet
'elm-trees' w ill almost certainly have a 'weaker' ste reot ype than
' t i g e r s ' . *4 One can consider t h a t e i t h e r the d e c i s i v e f a c t o r i s the
taxonomic l e v e l o f c at e gor i es (e.g. 'animal' may be a compar ati vel y
hi gher l e v e l category than 'tre e ' and on the same l e v e l as ' p l a n t ' ,
h e n c e ' t i g e r ' may be hi gher than ' e l m - t r e e ' ) or t h a t r e l a t i v e sal i ence
can only be determined i n connection w it h the perceptually nearest
t r e e s or animals i n such a case. The important thi ng seems to be the
recognition of such thi ngs as the relative salience of different
p r o p e r t i e s o f an o b j e c t , which Putnam does not seem t o be t a k i n g i n t o
c o n si d e r a t i o n .

Another problem for t he theor y is that the s te r eo t yp e is not

restricted as t o t he type o f i n for ma ti on it may i n c l u d e , so i t can
boi l down to eve ry thi ng we ass oc iate wi th a word that helps us
understand i t . Ways o f coping wi th these problems w i l l be discussed
in the secti on on Rosch's pr otot ype s. One t h i n g , however, has to be
pointed out here. There are two p i t f a l l s we wish t o avoid: one of
them is a ll ow ing the s te reot ype to i nclude e xp e rt knowledge f o r every
term, i.e. i n for ma ti on constituting necessary and sufficient
c ondi ti ons f o r membership in the ext ensi on; the ot he r one i s a ll owi ng
it to i nclude p u r e l y individualistic information. It is c l e a r from
what has preceded t h a t the f i r s t caveat is taken care o f ( a t l e a s t in
theory). In Putnam's own words:

"A world in which everyone i s an expert on every t o p i c i s a world

in which soci al laws are almost unimaginably d i f f e r e n t from what
they now a r e . What i s the m o ti v at io n o f t a k i n g such a world and
such a language as the model f o r the a n al y s i s o f human language?"
( i b i d . :187).

It is not equally clear, perhaps, how the second danger w ill be

avoided. The s te r eot y pe is said to include i n for ma ti on about the

- 56 -
minimum s k i l l s r e q ui r ed fo r entry i n to a linguistic community (what
Putnam c a l l s " s ig n if ic a n t information").

While necessary and sufficient c o n d it io ns , or knowledge of the

'essential' characteristics may be a r r i v e d at (for some terms, at
least) w i t ho ut having recourse to the judgments of native speakers
(but only to the experts), 's u p erficial' or stereotypical
characteristics r e q u i r e compl etel y d i f f e r e n t methods. Introspection
w ill not suffice. The importance of e lic itin g i nf or ma ti on from
native speakers becomes i m p e r a t i v e . Putnam's t he or y does not r a i s e
these issues directly. My own interpretation of the Stereotype
approach to meaning and i t s possi bl e p r a c t i c a l i m p l i c a t i o n s consi sts
o f the f o l l o w i n g main p oi nt s:

1, Obviously some s t e r e o t y p i c a l i nfor mati on w i l l n e c e s s a r i l y be

encyclopaedic or empirical; some of it w ill be also
incorrect; but the p o s i t i o n : "whatever is revisable cannot
be linguistic but simply encyclopaedic, empirical,
scientific", which is stifling for lexical semantics, is
refuted.

2. Extremely subjective m at e r i a l (which can be detected in

answers t o l e x i c a l tests) can be shown not to cover common
ground and therefore stay at the bottom as 'm arginal':
either le ft out o f an attempted normal form d e s c r i p t i o n or
included as opti onal ( n ot i c e that Putnam distinguishes
between ' o b l i g a t o r y ' and ' o p t i o n a l ' stereotypical fe a tu re s ).

The exact na tur e o f the s t e r e o t y p i c a l i nf or ma ti on we w i l l arrive at

can only be discussed ( i n connection wi th the s p e c i f i c terms we are
investigating) after a sufficient amount of material has been
accumulated. For t he moment we seem to have more knowledge o f what
the s t er eot ype should not contai n and much l ess on what it should.
What we further know on the basis of Putnam's t he or y is that a
distinction is to be drawn not only between stereotypical and
'essential' characteristics but also between these and semantic
markers. Not ic e firs t, that semantic markers i n vo lv e i nf or mati on

- 57 -
which i s " c e n t r a l " t o the term, " p a r t of a w i de ly used and important
system of classification" ( i b i d . :189). A semantic marker is
therefore a hi gher level pr op e rt y (in the taxonomic sense), i.e. a
more i n c l u s i v e one ( e . g . 'liq u id ', 'animal', e tc.).

The most i mportant pr ope rt y o f semantic markers f o r the d e s c r i p t i o n

o f terms ot he r than those Putnam has analysed ( e . g . motion verbs) is
precisely that they function as classifying units. This is
r emi ni sc ent of the traditional distinction between major and minor
features and seems to me v a l i d for the semantics of MGMVs. How
e x a c t l y major c l a s s i f y i n g features f o r MGMVs d i f f e r from minor ones
w ill h o p e f u l l y become e v id e nt in the course o f the present discussion
o f these verbs. It should be poi nted out here, though, that very
little i s known about the kind o f n a m ab i l i t y p r i n c i p l e s which are at
work in the domain of verbs in general and that arriving at an
overall specification of such principles lies clearly outsi de the
scope o f the present i n v e s t i g a t i o n . We do, however, know t h a t we are
d e al i ng w it h rather abstract ont ol o gi c al c a t eg or i e s such as
intentionality and the related noti on of agentivity, for i nstance.
In the domain o f motion verbs what are u s u al l y understood as high-
level properties are l i n k e d wi th the f ol l o w i n g categorizations: (a)
s t a te s vs processes vs events and (b) a g e n t i v i t y and c a u s a t i v i t y ( i n
relation to a c t i o n s ) . There can be l i t t l e doubt t h a t such f e at u r e s
are 'maj or c l a s s i f y i n g ' ones in Putnam's sense.

It w ill become obvious in the course of the di scussion of these

features that I understand them as s ynt a ct ic o- s em ant ic in na tur e.
Now the distinction between mass and count Ns is allotted to the
' s y n t a c t i c markers' box by Putnam. It w ill be shown, however, that
although one candraw a p a r a l l e l between t h i s distinction and the
'event-process' one (Leech 1971, Mourelatos 1978) the issue i s much
more complicated when it comes to Vs, the whole phenomenon being
graded r a t h e r than exhibiting a simple dichotomy, much l ess clear-
c u t, depending on s y n t ac t i c o- s em a nt i c frameworks and some pur el y
semantic factors r e q u i r i n g speci al analysis. So, to start with,
compared t o the sharp d i s t i n c t i o n Putnam draws between s y n t a c t i c and
semantic markers in his normal form d e s c r i p t i o n s for n a tu ra l kind
terms, i t would seem more a p p r op ri at e to p os t u l a t e a u n i f i e d category

- 58 -
of s y n t ac t i c o- s em a nt i c markers to accommodate ' maj or classifying'
fe atures. It should be f u r t h e r noti ced t h a t no ' e x t en s i o n box' type
o f i n for ma ti on can p os si b l y go i n t o the semantic d e s c r i p t i o n o f MGMVs
at l e a s t . This i s not due t o 'absence o f speci al knowledge', because
in the strict sense of the term, one can easily think of such
knowledge being possessed by an a t h l e t i c s committee, for i nst an ce ,
concerning the d i s t i n c t i o n between ' r u n ' and ' w a l k ' . It is ju s t that
no such bit of knowledge can amount to necessary and sufficient
c ondi ti ons for isolating any motion verb from any o t h e r one in the
'practical' sense a l r e a d y discussed and in the more t h e o r e t i c a l one
presented in t h i s s ecti on concerning the competence o f the non-expert
speaker.

1.3.2 On Rosch's prototypes

1.3. 2.3. Stereotypes and pr otot ypes: common ground

Psycholinguistic research usually associated with E. Rosch and C.
Mervis has also provided evidence that the meaning of words is
inherently f uzzy and i ndet ermi na te and t h a t communication succeeds
because we us u a l l y operat e w it h protot ypes. If categories^ were
defi ned i n mental representation by a set of c r i t e r i a (properties/
attributes) constituting necessary and sufficient c ondi ti ons for
membership, we would expect a l l members to be e q u a l l y r e p r e s e n t a t i v e
o f the cat egor y. Rosch and her colleagues have shown however t h a t
some members are more characteristic of the category than other s,
i.e. more prototypical. Their hypothesis is that c at e go ri e s are
maintained as d i s c r e t e by being coded in c o g ni t i o n in terms o f the
prototypes o f the most c h a r a c t e r i s t i c members o f the c at egor y. Their
experiments have produced evidence in favour of this hypothesis
showing t h a t c a t eg or i e s are coded in the mind by means o f a prototype
of a ty pical category member, i.e. a 'c o nc r et e image' o f an average
category member. (Rosch 1977b:2 13 -4 , Mervis and Rosch 1981).

The basis of the common ground covered by Putnam's and Rosch's

th e o r i e s seems t o be t h a t c at e go r ie s are not s p e c i f i e d by necessary
and s u f f i c i e n t c ondi ti ons f o r membership and t h a t their boundaries
are not wel l defined. A reasonable combination o f the two is t h a t

- 59 -
the s t er eot ype could be understood as a c l u s t e r o f a t t r i b u t e s o f the
most p r o t o t y p i c a l member(s) o f a category, and t h a t word meaning can
only be given through an i m per f ec t and open-ended d e f i n i t i o n of the
p rot ot yp e.

Putnam i s bothered by what he c a l l s "the idealization o f supposing

that there is such a t hi n g as the set o f t hi ngs o f which the term
'tre e ' i s t r u e " . 16 Ful l r e co g n i t i o n o f the f a c t t h a t membership in
the extensi on i s not n e c e s s a r i l y a ' ye s- no' question i s also given by
the pr otot ype approach which a c t u a l l y provides the means, or a t l e a s t
attempts at working out a method f o r measuring degree o f membership
through a d i r e c t appeal to n a t i v e s p e a k e r - 1 i s t e n e r s . This way, such
facts, i nstead o f being only a piece o f phi losop hi ca l th eorizing (in
the a b s t r a c t ) , can become the o bj e c t o f d e t a i l e d a n al y s i s and l i n k ed
to the ' average speaker-1istener' whose intuitions may be either
ignored or merely guessed at by phi l osophy- of-l anguage p r a c t i t i o n e r s ,
but not by descriptive linguists who purport to try and account
p r e c i s e l y f o r these very i n t u i t i o n s .

Both approaches (Putnam's and Rosch's) recognize the importance of

the r o l e played by the actual nature o f obj ect s i n the semantics of
terms and t h a t d i f f e r e n t kinds o f terms correspond to d i f f e r e n t kinds
o f competence. This i s in i t s e l f q u i t e s i g n i f i c a n t as i t may be used
to predict that all combinations of features are not e q ual l y
probabl e. (Other theories cannot exclude the conj unct ion of all
possi bl e a t t r i b u t e s . ) Rosch observes t h a t the percei ved world does
not contai n random c l u s t e r i n g s o f a t t r i b u t e s . Some combinations are
more expected than others and some are completely i mpossi bl e. So,
for i n st an ce , while 'feathers', 'fu r' and 'wi ngs' are separate
attributes, 'feathers' are more l i k e l y t o co-occur w i t h 'wi ngs' than
'fu r' is. This is an e mpi ri ca l f a c t perceived in the r e al world and
i s only compatible w i t h an e x t e n s i o n a l i s t semantic t he or y such as the
one Putnam o u t ! i n e d .

Putnam does not s p e c i f y a t a l l what kind o f i n for ma ti on goes i n t o the

s te reot ype o f terms ot he r than n a tu ra l kind ones (as a lr ea dy pointed
out), w h i le Rosch maint ai ns that the i nfor ma ti on constituting a

- 60 -
protot ype consi sts b a s i c a l l y of perceptual and f u n c t i o n a l properties
(e.g. 'legs' and 's it-o n -a b le n e s s ' as attributes of 'chairs' are
examples of a perceptual and a f un c t i o n a l pr oper ty respectively).
Notice, however, that such properties are only specified in
connection wi th concrete nouns and wi th r espect t o a p a r t i c u l a r l ev el
of abstraction which Rosch calls 'basic level' and which w ill be
given speci al attention in what follows. L ittle , if anythi ng, is
said by e i t h e r t he or y regardi ng the nature o f a t t r i b u t e s of anything
ot he r than concrete nouns.

1.3.2.2 Prototype t heor y as a theor y o f c a t e g o r i z a t i o n

I have so f a r concentrated only on how Rosch's approach r e i n f o r c e s
the i dea, through p r o t o t y p i c a l i t y t e s t s , t h a t t h e r e i s no such thi ng
as necessary and s u f f i c i e n t c ondi ti ons f o r membership. That subjects
could make judgments concerning degree o f membership o f an item in a
cat egory, thus accepting t h a t ' o l i v e s ' , ' p i n e a p p l e s ' and ' a p p l e s ' are
not equally typical of the category 'fru it', for i n s t an c e, was
r epor te d by Rosch as e a r l y as 1973. The i m p l i c a t i o n is that i f the
alternative hypotheses were correct (those requiring necessary and
sufficient c o nd i t i o n s ) sim ilar answers would be elic ite d for all
members of an assumed superordinate/inclusive category. This
difference in di st an c e is computed within a given system, i.e.
presupposes (as a lr e ad y mentioned) a p r e vi o us l y e s t a b l i s h e d taxonomy;
but the r e l a t i o n s h i p between i n c l u s i v e category and i ncluded members
is not simply a relationship of inclusion. The category which
c onsi sts of a pr otot ype or representative i nstance and ot he r less
representative or marginal instances around it is understood as
i n t e r n a l l y s t ru c t u r e d i n t h i s p a r t i c u l a r way.

Rosch e t a l . (1976:383) recognize w i t h i n concrete noun taxonomies one

particular level, the 'basic level of abstraction', at which
c at e gor i es carry the most i nfor ma ti on and are maximally different
from one another. Two more l e v e l s are r ecognized, p a r a l l e l t o those
e s t a bl i s h e d f o r b i o l o g i c a l taxonomies, f o r i n st an ce , a most i n c l u s i v e
one (above the basic l e v e l ) c onta ini ng s uperor di nate c a t eg or i e s ( e . g .
'fru it') and a l e v e l lower than the basic one c o n ta i ni n g subordinate

- 61 -
c a t eg or i es (e .g. golden delicious' for the basic level category
'a p p le ').

1.3.2.3 The basic l e v e l o f a b s t r a c t i o n

The basic l e v e l of abstraction is claimed to be the primary l e v e l ,
where 'cuts' are made in the continua of the environment, and
constitutes the level "a t which the organism can o bt a in the most
i nf or ma ti on wi th the l e a s t c o g n i t i v e e f f o r t " (Rosch 1977b: 2 1 3 ) ; hence
its identification i s o f extreme importance f o r t h e o r e t i c a l reasons,
but also quite necessary even for the practical purpose of
e s t a b l i s h i n g hyponymic r e l a t i o n s .

Basic level words were firs t identified by Rosch on the basis of

questions like 'What is this?' w h i le showing t o subj ects an apple,
for i n st an ce . The response was expectedly 'an a p p l e ' , i.e. neither
the super or di nate (fru it) nor the subordinate (a golden d e l i c i o u s ) .
This is e xpl ai ned on the basis of its r e pr es e n t i n g a 'cognitively
effic ie n t' level, at which the i nf or ma ti on value of attribute
clusters is maximized (hence it is posited as the most fundamental
level o f a taxonomy). Mervis and Rosch ( 1 9 81 : 92 ) cite a number of
studi es in d i f f e r e n t domains supporting the e xi s te nce o f such a l e v e l
(e .g. N. Cantor e t a l ., f o r p s y c h i a t r i c c a t eg or i es and B. B e r l i n e t
1976,
a l. 1973 and C.H. Brown et al.,^ w ixh linguistic and cultural
e vi d enc e) . Psychological r esear ch, in particular, suggests that
obj ect s are recognized as belonging t o basic level c a t eg or i e s more
r a p i d l y than as members o f c at e go ri e s a t o the r l e v e l s (Rosch e t a l .
1976) and t h a t they are those l i k e l y to be l ea rn ed f i r s t by the c h i l d
( B e r l i n e t a l . 1973).

Let us consider b r i e f l y whether the c h a r a c t e r i s t i c s of 'spontaneous

naming' and ' maxi mi zati on of i n f o r m a t i o n ' 17 are also applicable to
ot he r areas o f the vocabulary and in p a r t i c u l a r t o MGMVs.

If one poi nts a t somebody walking in the s t r e e t ( i n a very o r di n ar y

way) and asks 'What i s he d o i n g ? ' , duplicating Rosch's question f o r
na t ur a l kind c a t eg or i e s and a r t e f a c t s , the response i s more l i k e l y to

- 62 -
be 'p e rp a ta i' (he is wal ki ng) than 'v im tiz i' (he is pacing,
steppi ng) or ' siya n o p e rp a ta i' (he is walking slowly), which are
hyponyms o f 'p e rp a ta o ' (walk) - provided, o f course, one is asking
about a ( v e r y) good i nstance o f the cat egor y. This i s a l l very wel l
provided we all agree that 'perpatao' (wal k) is a basic level
cat egor y. The problem i s t h a t it i s ext remel y d i f f i c u l t , if at a l l
possible, to d i scov er a super or di nate category which can be s a f e l y
said t o include verbs such as ' p e r p a t a o ' and ' t r e x o ' (run). The only
candidate seems to be 'p iy e n o ' (go) . Not ice an immediate
compl icati on: i f the goal i s obvious, e . g . someone wal ki ng to a n e a r ­
by kiosk , the answer (as f a r as I have checked) is: 'p iy e n i sto

p e rip te ro ' (he is going to the kiosk ) and not 'p e rp a ta i (pros to
p e rip te ro )' (he is walking (to the kiosk )). In short, a lot
depends on what is assumed to be common knowledge. Context,
pr esupposi ti on and predominantly pragmatic reasons are i nvolved in
identifying the var ious levels of abstraction on the basis of
'Question and Answer' t a sks . This has been most c o nv in c i n g l y shown
by Cruse (1977) for nouns. E la bo r at in g on the e x t r a compl icati ons
f o r Vs seems unnecessary here. The p a r t i c u l a r problems o f verbs w i l l
be discussed' at many different poi nts in the course of this
investigation.

Besides 'spontaneous naming' the basic level is also ascri bed the
characteristic of maximization of i n f or m a t i on . Evidence for this
claim is provided by the experiments mentioned in Rosch e t a l . (1976)
where subj ects were asked to lis t attributes for c at e gor i es
d i s t r i b u t e d over a l l t h r e e taxonomic l e v e l s . Counting the number of
a t t r i b u t e s common t o c at e go ri e s a t the thr ee l e v e l s the experimenters
found that basic level c at e gor i es had many more shared attributes
than super or di nate ones (e.g. for 'fru it', super or di na te : 3 shared
attributes; ba si c: 8.3 shared attributes). So the basi c level
appears as more i n f o r m a t i v e than the superordi nate one. Subordinate
c at e gor i es were found to have more shared a t t r i b u t e s than basi c l e v e l
ones, but the majority were shared wi th the basic level category
i n c lu di n g them and were t h e r e f o r e common t o most s ubor di nat es. This
impl ies l es s d i s c r i m i n a b i l i t y . So the subordinate l e v e l a ls o appears
as less i n f o r m a t i v e than the basic one.

- 63 -
These r e s u l t s are ext remel y i n t e r e s t i n g in themselves, although the
possibility of obtaining further results for o th e r kinds of
c at e gor i es in order to check the g e n e r a l i t y o f the p r i n c i p l e depends
crucially on being firs t able to obtain listings of attributes
d i r e c t l y from s u bj e c t s . The d i f f i c u l t y o f the task w i l l be discussed
in the next s ec t i o n .

1.3.2.4 Category formati on

There can be little doubt that the world consi sts of an infinite
number o f d i f f e r e n t s t i m u l i ; and t h a t one o f t he most basic f unc ti ons
of all organisms is c l a s s i f y i n g , i.e. d i v i d i n g up the world in order
t o cope wi th t h i s in fin ite diversity. I t has also become e vi de nt in
r ece nt y e ar s , at least, that this segmentation is not completely
arbitrary. It has a l re ady been mentioned in connection w it h
'colour', for i ns t an c e, that c at egor i es form around perceptually
salient points i n a domain; one can l e g i t i m a t e l y , I think, consider
such poi nts as 'cognitive p r o t o t yp e s' of the domain. Yet the
principles unde rl yi ng c a t e g o r i z a t i o n are s t i l l a t an e a r l y stage o f
r esear ch.

Rosch formul ates a hypothesis on the formation o f c at e go ri e s based on

a combination o f the two p r i n c i p l e s al re ady hi nte d a t in the previous
s ec ti ons : the principle of ' p e r c ei v ed world structure' (1978:29),
confirming that attributes co-occur in specific bundles, and the
principle of 'cognitive economy' (1978:28), which ensures that
c a t eg or i e s are formed so as to maximize discrim inability (and
therefore in fo rm a tio n ).

The important issue of how the inte rnal structure of c at e go r ie s

a r i s e s amounts, w i t h i n the framework o f t h i s th e or y , to establishing
the p r i n c i p l e s r e spons ibl e f o r the formation o f category prototypes
and g r adi ent s o f category membership, and has t o be presented in some
detail. For c at e g o r i e s w it h a p hy si o l o gi c al basis (e.g. col our s)
prototypes may be s t i m u l i which are s a l i e n t p r i o r to f or mati on of the
category and the very salience o f which determines the categorial
structure of such domains. For most ot he r domains prototypes are
understood as being formed through principles of learning and

- 64 -
inf or ma ti on processing from the items given in the cat egory (Rosch
and Mervis 1 97 5 : 5 7 4 ) . One o f the major s t r u c t u r a l p r i n c i p l e s which
are said to govern t he formati on of the prot ot ype structure of
semantic c a t eg or i es is that of 'family resembl ance', borrowed from
Wi tt g e n s t e i n (1953). It suggests t h a t it s u f f i c e s f o r each r e f e r e n t
o f a word to have a t l e a s t one pr ope rty in common w i t h one or more
ot he r referents, w h i le at the same time it is pos si bl e for few
p r o p e r t i e s or no p ro p e r t y at a l l t o be common t o a l l o f them.

Rosch and Mervis (1975) build on t h i s notion t h e i r hypothesis t h a t

members o f a category are understood as p r o t o t y p i c a l to the e xt ent
t h a t they bear a f a m i l y resemblance to other members o f the category
- i.e. the hi gher the degree o f f a m i l y resemblance, the g r e a t e r the
number o f p r o p e r t i e s shared w i t h ot he r members o f the same category.
Conversely, items viewed as most prototypical of a particular
category are those w i t h least family resemblance t o ( or membership
in) ot he r categories. There is a strong correlation between the
score an item gets on the basis o f p r o t o t y p i c a l i t y judgments and the
one i t gets on the basis o f i t s f a m i l y resemblance (number o f shared
attributes). In practical terms this means that a prototypical
member o f the category 'fru it' (e .g. 'apple') shares most a t t r i b u t e s
wi th o the r members o f the same category ( e . g . 'orange', ' p e ac h' ) and
fewest wi th members o f c a t eg or i e s ot he r than 'fru it' but on the same
level o f a b s t r a c t i o n w it h i t ( s upe ror di nat e l e v e l ) , e.g. 'fu rn itu re',
'musical instrument'. In this sense prototypical members of
c a t eg or i es have the p r o p e r t i e s o f maximal i nfor mati veness and maximal
discrim inability. The p r i n c i p l e o f maximization of i nf or ma ti on and
discrim inability is considered r esponsi bl e both f o r the s al i ence of
the basic l e v e l o f a b s t r a c t i o n and f o r the s al i e n c e o f protot ypes.

It i s obvi ousl y o f c e n t r a l importance to the present a n al y s is to be

abl e to t e s t the v a l i d i t y o f these p r i n c i p l e s f o r verbs. The major
o bstacl e is the one st at ed f o r the basic l e v e l o f abstraction: the
difficu lty of o bt a i n i n g lists of attributes (compiled by subj ects)
comparable to the ones Rosch and her colleagues have been able to
obtai n f o r concrete nouns. Far from being c h a r a c t e r i z a b l e in terms
o f ' pe rc ept ual and f u n c t i o n a l p r o p e r t i e s ' such as those presented f o r
the concrete nouns tested, most verbs i n vo lv e far more complex

- 65 -
characteristics. The problem is discussed in detail in Pulman
( 1 98 3) . Pulman reports the results of experiments he conducted
roughly duplicating the ones re por te d by Rosch but replacing noun
ca t egor i es w it h verb ones. Subjects were asked to w r i t e down as many
a t t r i b u t e s as they could t h i n k o f ( w i t h i n 90 sec. per item) as being
characteristic of the specific verbs they were presented wi th (or
r a t h e r o f i nstances o f the events or act ions these verbs stood f o r ) .
The responses ranged from synonyms and d e f i n i t i o n s t o connotati ons,
the category name i t s e l f and some a t t r i b u t e s o f the kind Rosch had
obtained f o r nouns (e.g. 'hug' = 'usi ng both a r m s ' ) . What is much
more s er i o u s , i f the a t t r i b u t e s obtained were t o be e d i t ed in the way
suggested by Rosch, over 50% of the original ones (for each item)
would have t o be di scounted as c l e a r l y i n a p p r o p r i a t e . In view of a l l
this, i t can be no s u r pr is e t h a t f a m i l y resemblance was not found to
be p o s i t i v e l y correlated w it h prototypicality (which was precisely
the hypothesis put to the t e s t ) . Pulman's conclusion is t h a t f a m i l y
resemblance cannot be t h e r e f o r e said to c o n s t i t u t e a causal f a c t o r in
the formati on o f protot ypes (Pulman 1 9 8 3 : 1 1 9 - 2 0 ) .

Two thi ngs have to be noti ced here. Firstly, obt ai ni ng

prototypicality judgments for the same sets of verbs and their
respective subordinates was no problem f o r Pulman. Such judgments
were also easy to obta in from Greek subj ects ope rat in g wi th MGMVs,
The actual details of these t e s t s w i l l be discussed in the r e l e v a n t
c hapt er . It seems, however, in order to o u t l i n e r i g h t a t the outset
what the t h e o r e t i c a l possibilities are (given the present state of
t he art). The question is not so much whether the p r o t o t y p i c a l i t y
effect holds for verbs, but rather which principle(s) is(are)
a c t u a l l y at work t h e r e . This leads us to the second p o i n t . Leaving
subj ects a s i d e , f o r a moment, and the s o r t o f i n f or m at io n we normally
ass oc iate w i t h verbs, consi der some compl icati ons which a r i s e before
we start l ooking for attributes. Take the classical example of
'k ill' which is probably a clear case of a basic level term and
relatively easy to find subordinates for. A po s si b le set of
subordinates would i nclude 'murder', 'execute', 'a s sa s sin a te ', etc.
An e q u a l l y p l a u s i b l e set would consi st o f ' s t a b ' , ' run ov er ' (wi th a
car), 'stran g le', etc. The members of the former set are
d i f f e r e n t i a t e d in terms o f 'purpose' or ' m o t i v e ' , those o f the l a t t e r

- 65 -
one in terms of 'manner'. This gives rise to a number of
compl icati ons which run counter to the very idea of testing the
principle of 'fam ily resemblance': firs tly there are problems of
category i n cl u s i o n which are not comparable to those o f nouns, A
stabbing or a s t r a n g l i n g event might be reasonably considered an
i nstance of ' m u r d e r i n g ' . This does not, however, necessarily imply
that 'murder' is to be considered a s uper or di nate of 's tab' or
'strangle' and t h e r e f o r e hi gher than or a t the same l e v e l as ' k i l l ' .
For while 'murder' necessarily i m pl ie s 'k ill' (and is therefore
included in the cat egor y ' k i l l ' ) , 'stab' and ' run o v er ' may not
result in a murdering or killin g event at a l l , c o n tr a ry to
's tra n g le '. Such problems o f c l a s s - i n c l u s i o n are not l i k e l y to turn
up wi th the kinds o f nouns Rosch and her col leagues st udi ed.

Consider now the situation wi th respect to specific attributes for

these terms and t he formati on of protot ypes. The members o f the
firs t s et o f hypothesized hyponyms o f 'k ill' share the proper ty of
' intentionality' unlike the category name i t s e l f which is unmarked
for this fe ature. In addition to this, they all share what they
inherit from the category name, i.e. 'cause-to-die' or something to
that effect. They are d i f f e r e n t i a t e d in terms o f d i f f e r e n t motives,
the evaluation of which cannot be expected to be unanimous:
'executing' may be a l awf ul a ct ion but f o r some speakers a t l e a s t i t
may wel l carry the connotations associated wi th 'murdering'. The
second s et o f hypothesized hyponyms contains only one term which is
p ro p e r l y included i n the hi gher c l a s s , namely ' s t r a n g l e ' and a number
o f others which are not p ro p e r l y i ncluded, i.e. 'run o v e r ' , 'stab',
'shoot', etc. One term w ill stand out as uns p ec if ie d for
'in te n tio n a lity ', namely 'r un over'. They are differentiated in
terms o f d i f f e r e n t 'means' through which the r e s p e c t i v e actions are
carried out, e.g. 'w ith a car' for 'run over', 'with a gun' for
'shoot', 'w ith a knife' for 'stab', 'through pressure exercised by
hands, e tc.' for 'stran g le'. By the sheer f a c t that ' r u n over ' is
s ingled out for not being specified for ' intentional i t y ' it will
probably receive a lower fa mily resemblance score than 'stab' for
i ns tan ce . The only cat egory p ro p e r l y i ncluded in t he hi gher one
(irrespective of whether 'k ill' or 'murder' is considered the
immediately hi gher cat egor y in this c as e ), namely 'stran g le', w ill

- 67 -
 probably r e c e i v e a lower f a m i l y resemblance score than the r e s t , for
the very reason which renders i t the most ' l e g i t i m a t e ' subordinate o f
'k ill' or 'murder', namely t h a t it normally involves 'cause-to-die'
(i.e . n e c e s s a r i l y i mpl ie s ' k i l l i n g ' ) .

At this preliminary stage

k these observati ons only touch on the problems i nvolved and are simply
meant to show the messiness o f the data o f a t t r i b u t e s f o r verbs and
t he fu tility of counting attributes and establishing family
resemblance as a p l a u s i b l e expl ana ti on of cat egory formation f o r such
categories. If 'frequency' and c u l t u r a l 'salience', (mentioned in
Rosch and Mervis 1975:599 as f a c t o r s also c o n t r i b u t i n g to category
formati on and in p a r t i c u l a r to the formation o f protot ypes) are also
considered very b r i e f l y in connection wi th the above d a ta , t he r e s u l t
is not much better for the family resemblance principle. The
assumption i s t h a t the most f a m i l i a r obj ects are named f i r s t and are
basic level (e.g. cats, chairs). Superordinates are then formed
around such basic l e v e l terms which are alr eady e s t ab l i s he d and which
become the protot ypes of the superordinate (e.g. of 'animal' and
'fu rn itu re' respectively). It seems in tuitively correct that
compared to ' s t a b ' , 'r un over ' is more fam iliar and somehow more
culturally salient and more f re que nt (in terms of 'frequency of
contact'), yet their respective family resemblance scores are not
very l i k e l y to account f o r t h i s intuition.

Fam iliarity and relative salience, although in bad need of an

accurate e x p la n at io n or a t l e a s t of some e l a b o r a t i o n , look l i k e more
promising f a c t o r s in the case o f a b s t ra ct c at e go r ie s such as those o f
verbs. They w i l l therefore be looked at in c onsi de ra bl e d e t a i l in
the chapt er di scussing p r o t o t y p i c a l i t y t e s t s f o r MGMVs.

I t has to be poi nted out once again t h a t t h e r e i s no reason t o expect

semantic competence f o r verbs to be sim ilar to that for nouns.
Moreover, i t is plausible t o expect d i f f e r e n t a t t r i b u t e s in d i f f e r e n t
verb domains. In general what subjects seem to grasp as verb
attributes can be shown to be more 'vague' and 'ina cc ura te ' (in
Putnam's sense) than the 'perceptual' and ' f u n c t i o n a l ' properties of
the nouns Rosch and her col leagues i n v e s t i g a t e d , even w i t h i n domains

*( F o r detailed discussion see chapters 4 and 5).

- 68 -
involving relatively 'concrete' m a t er i a l such as that of motion
verbs. None o f these c on si d er a ti o ns invalidates Rosch's hypothesis
f o r category formati on in the s p e c i f i c areas she has examined. The
observati ons made in t h i s s ecti on do not bear on the issue o f whether
certain events are more p r o t o t y p i c a l of an i n c l u s i v e event or not.
They simply r a i s e questions concerning class i n c l u s i o n , the v a l i d i t y
o f the ' f a m i l y resemblance' p r i n c i p l e and the f e a s i b i l i t y of the task
of arriving a t comparable (and t h e r e f o r e also countabl e) attributes
by d i r e c t appeal to n a t i v e speakers in the case o f verbs ( or r a t h e r
in domains other than the ones Rosch has been concerned w i t h ) .

What seems to be completely i n d i s p e n s i b l e in Rosch's theor y i s t h a t

c a t eg or i e s are internally s t ru c t u r e d by g ra d i e n t s of
r ep re se n ta t iv en e ss and category boundaries are not definite. The
principles of cat egory formati on for different types o f c at egor ies
can probably be best explored after c onsi de ra bl y more experimental
work has been done in different areas, checking firs t whether the
protot ype e f f e c t i s o f as wide a p p l i c a b i l i t y as i t i s expected to be
(on the basis of theoretical and experimental work done so far).
Most of Rosch's more r ece nt work concentrates on 'degree of
membership' rather than the nature of the protot ype as a mental
construct. So definite answers on the principles governing its
formati on are r a t h e r premature. Yet even as a piece o f e x pl o r a t or y
work, if not a complete theor y of categorization, Rosch's
investigation raises a number o f s t i m u l a t i n g questions and suggests
some possi bl e answers. Besides, as i t depends h e a v i l y on speakers'
judgments, it attracts attention to the tremendous significance of
t e s t s even f o r d e s c r i p t i v e l i n g u i s t i c purposes.

It i s possi bl e t h a t t e s t s d i f f e r e n t from the ones Rosch has used are

more a p p r o p r i at e f o r ot he r domains (such as the semantic s i m i l a r i t y
t e s t s discussed here in the r e l e v a n t c h a p t e r ) . Although expl anati on
o f experimental r e s u l t s may be ' a poor s u b s t i t u t e ' f o r understanding
human m e n t a l i t y (Johnson-Laird and Wason 1977:2), it seems quite
i ndi spensable for a semantic d e s c r i p t i o n which wishes to take i nt o
c ons id e ra ti o n ' o t h e r people and the w o r l d ' .

- 69 -
1.4 D e l i m i t a t i o n o f the f i e l d o f motion verbs

It is wel l known t h a t a l i s t c on ta i ni ng a l l and only verbs o f motion

i s "an impossible i d e a l " ( M i l l e r and Johnson-Laird 1 9 7 6: 5 30 ) . It is
not in e f f e c t obvious t h a t such an ' i d e a l ' should be posi ted in the
firs t place, given the nature o f language and the fu zz i ne s s of the
boundaries of semantic fields, which was recognized long before
Prototype t he or y: Schweidenweiler (1942), among ot h e r s , already
n ot ic e s t h a t the boundaries o f semantic f i e l d s are not c l e a r .

I t seems p l a u s i b l e to accept t h a t the c l e a r e s t case o f a motion V i s

one which describes the change o f l o c a t i o n of an o b j e c t . This means
t h a t what i s common to a l l the Vs we wish w i t hout h e s i t a t i o n to c a l l
motion Vs i s t h a t they descri be (among ot he r t h in gs ) the f a c t t h a t : a
certain o bj e c t A which at a given time t was a t some po i nt x is a t
p o i n t y at a l a t e r time t ' . So, M a r i e t t a came home i mpl ies t h a t the
subj ec t was away from home a t time t - 1 and t h a t she was a t home at
some l a t e r time t .

In a d d i t i o n to the class o f Vs d esc ri bi ng a c h a n g e - o f - l o c a t i o n of an

object, verbs i n v o l v i n g a c h a n g e - o f - p o s i t i o n not o f the o bj e c t as a
whole but of p ar ts of it may be also understood as MVs. This
category might include Vs desc ri bi ng change of shape, orientation/
rotation or oscillatio n. So, in Marietta bent down, the motion
described by the V does not i nvolve a change-of-location of
'M arietta'. The change i n the subject's state described here is a
change in shape r e s u l t i n g from her going from the standing p o s i t i o n ,
for i n st an ce , to the one referred to by 'bend'. Motion but no
c h a n g e - o f - l o c a t i o n of the whole o bj e ct is i n vol ve d.

Some l i n g u i s t s , such as Ikegami (1969) and M i l l e r (1972), use the

term 'V o f motion' to r e f e r onl y to Vs which de scri be a change-of-
location ( o f the whole o b j e c t ) . Others, like Nida (1975), consider
as motion Vs also those r e f e r r i n g to c at e go r ie s such as the ones j u s t
discussed above ( wi th ou t r ecogni zi ng them as different c at egor i es
either). Nida r e f e r s to such Vs in English as i n v o l v i n g motion o f
' maj or parts of the body ', which he c o nt ra st s with Vs de sc ri bi ng

- 70 -
motion o f 'minor p a r t s o f the body', such as 'wink', 'sm ile', e t c.
In his a na ly s is it is hard to see: (a) how one would d i s t i n g u i s h
' m a j or ' from ' mi nor ' p ar ts ( consi der f o r i nstance t he case o f 'wa ve ')
and (b) how he excludes Vs de sc r ib i ng activities of vari ous sorts
such as ' w r i t e ' , 'c u t', 'p ie rce ', etc., since a l l activities i nvolve
motion o f some p a r t o f the o b j e c t performing them. No d i s t i n c t i v e
c r i t e r i a are o f f e r e d . Besides, no expl ana ti on i s given f o r excluding
s o - c a l l e d ' c o n t a c t Vs' such as: 'touch', 'h it', 'k ic k ', etc.

In •wechanics the motion of a body is analysed in terms of a

'translation' (change-of-location) of i t s cent r e of gravity, on the
one hand, and a r o t a t i o n or change o f shape or o s c i l l a t i o n , on the
o t he r . The l a t t e r type is r e l e v a n t as regards the motion o f par ts of
the body relative to the c ent r e of gravity or the position of
equilibrium (considered as a p oi n t of reference). This d i s t i n c t i o n
between c h a n g e - o f - l o c a t i o n and a l l ot he r kinds o f motion may be said
to be r e f l e c t e d in language at least to the extent that radically
different frames are compatible wi th motion verbs which f a l l into a
' c h a n g e - o f - l o c a t i o n ' category and those which do not.

Two major c at egor i es are t h e r e f o r e recognized: the firs t category

contains verbs d esc ri bi ng t he t r a n s f e r of an o b j e c t from one place to
another. All the Vs t h a t belong here are change-of-location
(hencef or th CL) verbs and they give rise to a contradition if
i n s er t e d in a frame l ike.:

(A) A _ _ -ed ( B) , but (B) did not change l o c a t i o n / p l a c e .

Examples of such verbs are 'go', 'tra v e l', 'e n te r', 'wander ',
'bring', 's end', etc.

The second category comprises verbs a l re ady r e f e r r e d t o as descri bi ng

kinds o f motion i n v ol v i ng pa rts o f the o b j e c t changing l o c a t i o n but
not the o b j e c t as a whole. These verbs are change-of-position
( he ncef or th CP) verbs and g i ve r i s e t o a c o n t r a d i c t i o n i f i n s e r t e d in
a frame l i k e :

- 71 -
 (B) A -ed ( B ) , but (B) did not move.

Examples of such verbs are 'trem ble', 'tu rn ', 'lean ', 's it down',
etc.

It is obvious that a great number of Vs d e sc r i b i ng a ctivities of

vari ous sor ts w i l l d i s p l a y the same behaviour concerning frame (B) as
the verbs j u s t discussed, e . g . :

(3) * t h e man wrote/was w r i t i n g , but he was not moving.

The same holds f o r 'p ierce', 'c u t', 'break' and a l o t more. What is
b a s i c a l l y described here i s the r e s u l t t h a t the motion o f the subj ect
has ( a l so on the object, if there is one), rather than the actual
motion as such. The a ct io n of'w ritin g ', for i nst an ce , can be
performed by hand, by mouth or by foot even (in the case of
handicapped people), and by machine; the point is, however, that
mimicking the movements t h a t remind one o f w r i t i n g w i t hout anything
r e s u l t i n g from i t , w ill not be r e f e r r e d to as ' w r i t i n g ' . So none o f
the Vs d e sc r ibi ng a c t i v i t i e s o f vari ous sor ts but not motion as such
need be considered motion Vs. The same could be said to apply to
what M i l l e r (1972) calls 'contact' Vs. The category would comprise
Vs such as: 'h it', 'be a t', 'push', 'slap ', 'touch'.

The issue o f what constitutes a verb o f motion and what does not
seems to present problems even to those l i n g u i s t s who equate motion
w i th change-of-location of the whole object. Thus Ikegami (1969)
considers a verb as a verb o f motion i f the f e a t u r e o f motion is the
principal and not a subordi nate element of its meaning. He also
allows f o r the p o s s i b i l i t y t h a t a verb may cease to be a motion verb
"when we can no l onge r assign the f e a t u r e o f movement t o i t " (p.87),
and mentions ' e s c a p e ' , 'v is it' and ' e l o p e ' as examples o f verbs where
features ot he r than motion ones have become so prominent as to
overshadow the movement f e a t u r e . 'Assemble' is quoted as an example
o f a V which "ceases" to be a motion verb in We are assembled vs We
have assembled (where, he would cl ai m, the motional uses are f e l t to
be o r i g i n a l and the non-motional ones d e r i v a t i v e ) . Besides, Ikegami

- 72 -
holds t h a t t he rever se process is also p o s s i b l e , in the sense t h a t
the f e a t u r e o f movement may s t a r t o f f as a subordi nate element (or
even n o n - e x i s t e n t ) and become prominent. On t h a t basis he analyses
'throw' in:

( 4) He threw the b a l l ski 1 f u l l y

as "make a c e r t a i n movement o f the arm" + "cause something to go

through the a i r " ( p . 89) w h i l e in ( 5 ) ,

(5 ) He threw the b a l l over the fence

he analyses it as "cause something to go through the air" + "by

making a c e r t a i n movement o f the arm" (ib id .).

F irst of a l l , it i s q u i t e unc lea r how Ikegami decides when a f e a t u r e

is prominent or primary and when i t i s s ubordi nat e. Besides, i t is
not c l e a r t h a t two d i f f e r e n t analyses have t o be proposed, unless a
system o f d e s c r i p t i o n is presented which provides the p r i n c i p l e s on
the basis o f which such d i s t i n c t i o n s are v a l i d . It i s p l a u s i b l e to
assume that different contexts w ill i nvolve different (to some
extent) understandings o f a verb. Once the sterotypical p ro p e r t i e s
of a verb have been established, the relative s al i e n c e of such
p r o p e r t i e s may be discussed in connection wi th s p e c i f i c instances o f
the event, a c t i o n , e t c . , descr ibed. In the absence of such l o g i c a l l y
prior steps, deci si ons on the appearance and disappearance of
f e a t u r e s seem unwarranted and probably premature. For the moment we
are c o n ce n tr a ti n g, best as we can, on whether the f e a t u r e o f motion
can be said t o belong to the meaning o f c e r t a i n Vs i n an i n t u i t i v e
r a t h e r than a t h e o r e t i c a l l y w e l l - e s t a b l i s h e d sense. This means t h a t
if a verb can be seen as d esc ri bi ng how an o b j e c t changes from a
place p a t time t t o another place p ' a t a l a t e r time t ' it i s a very
good candidate f o r the semantic f i e l d under c o n s i d e r a t i o n . At t h i s
stage nothing more can be said about ' t h ro w ' except t h a t in most o f
its uses as wel l as in i t s most c h a r a c t e r i s t i c instances i t seems to
i nv ol ve t h i s f e a t u r e . The 'most g e n e r a l ' and the 'most p r o t o t y p i c a l '
understanding seem to c oi nci de as regards the central pr ope rt y of

- 73 -
this field (i.e . 'motion') for the vast m ajo rity of the verbs
included here. Hence 'throw' does not seem to constitute a
b o r d e r l i n e case.

The s i t u a t i o n is more complicated in the case o f verbs such as 'push'

and ' p u l i ' . A sentence l i k e :

(6 ) He pushed the c a r t up the h i l l

is analysed by Ikegami as involving "'go' + 'a t the same time as

pushing'" (1 96 9: 8 9, 161) and on t h a t basis 'push' i s regarded as a V
o f motion. I assume t h a t in a d i f f e r e n t c o nt e x t , as f o r i nstance:

( 7) He was pushing the wall in vain

the V would be regarded as not being a motion ver b. It can be

argued, however, t h a t the i m p l i c a t i o n o f motion in (6) is the r e s u l t
of knowledge to th e effect that if an impulse is exercised on
something movable, provided c e r t a i n f a c t o r s are fav our abl e (e.g. the
impulse is suffic ien tly strong and t h er e is no obstacle), the
recipient o f the impulse moves. A further im plication is that one
pushes something if one's intention is to make it move away. In
terms of the most general understanding of 'push' it cannot be
considered a motion verb as such. Its prototypical image may,
however, be c l o s e r to the situation presented in ( 6) than the one
presented in ( 7 ) . This i s one o f the many instances which show t h a t
even within fa irly well-defined subfields, such as ' c h a n ge - of -
l o c a t i o n verbs' the borders are expectedl y f u z z y .

M iller and Johnson-Laird (1976), who also restrict the domain of

motion verbs t o those o f c h a n g e - o f - l o c a t i o n , consi der verbs such as
'breathe', 'shrug', ' c o ug h' , 'sm ile', 'sneeze', etc., as b o r d e r l i n e
cases, since no c h a n g e - o f - l o c a t i o n i s e x a c t l y i n vo l v e d . They apply a
test of the form He shrugged, but he changed location and He
shrugged, but he d i d n ' t change l o c a t i o n ( p . 529). Since both Ss are
acc ept abl e, the ' b u t ' t e s t excludes the set o f verbs i n question from

- 74 -
the f i e l d . They note, however, t h a t He shrugged h i s shoulders, but
they d i d n ' t change l o c a t i o n i s odd, which goes t o show t h a t motion i s
i nvolved somewhere. T h e i r conclusion i s t h a t i f the above set i s to
be i ncluded, then one should also include verbs such as ' o s c i l l a t e ' ,
'ro ta te ', 'spin', 'tu rn ', and sim ilar verbs describing rotary
changes. Besides, they consider t h a t verbs such as 'absorb',
'extend', 'fill', 'grow', 'wi den' and others implying 'change in
shape or s i z e ' would also have to be included ( i f 'breathe', 's hrug',
etc ., are not excluded) as i n v ol v i ng "movement of boundaries"
(ib id .).

Since all such deci si ons are in effect subjective and arbitrary,
nothing is illegitim ate. It seems, however, important t o identify,
even at such a preliminary stage, verbs which are fe lt to belong
together, i.e. which form a n at ur al class, on in tu itive grounds.
Evidently 's hr ug' constitutes a special problem, because causati ve
's hr ug' (e.g. 'he shrugged hi s s ho ul d er s ') does i n vol ve change-of-
l o c a t i o n o f the o b j e c t , w h i le intransitive ' s hr ug' does not i nvolve
transfer of the subj ec t as a whole. In t h i s sense it is slightly
mi sl eading to base deci si ons on the status of the n a t ur a l class o f ,
say, 'minor bodily-movement s p ec i f y i ng ver bs' on t h i s idiosyncratic
feature of a single member o f the s e t . ( N o ti ce t h a t this situation
does not a r i s e wi th any o f the remaining verbs M i l l e r and Johnson-
La i rd i nclude in the set in q u e s t i o n . )

A second remark has to be made concerning verbs o f ' r o t a r y motion'

('ro ta te ', 'tu rn ', e tc .). It seems t h a t a d eci si on as t o whether
they are to be included or not is independent of that for the
previous group. They c o n s t i t u t e a f a i r l y e a s i l y i d e n t i f i a b l e c la ss:
they seem t o i n vol ve some change o f o r i e n t a t i o n o f the whole o bj e c t
(as intransitives) probably w it h a fixed point.I f the field is
restricted to c h a n g e - o f - l o c a t i o n verbs, they are a b o r d e r l i n e case.
I f the f i e l d involves motion verbs in g e ne r a l , they are i n d i s p u t a b l e
candi dates.

The t h i r d set o f verbs mentioned in M i l l e r and Johnson-Laird includes

'absorb', 'extend', 'fill' and the l i k e and is said t o i nvolve

- 75 -
'movement of b o u n d a r i e s ' . Not ic e t h a t Ikegami (1969) considers verbs
such as these ( ' e x p a n d ' , 's tre tc h ', 'sw ell') as i n v ol v i ng p a r t of the
object moving w h i l e the rest remains in the same p l a c e . In this
understanding he lumps them t o g e t h e r w it h 'stand up', 'fa ll' and
'shake'. One might object to the i n c lu s i on of individual items in
either categorization but, as al re ady stated, the most important
t hi ng a t t h i s stage seems to be i d e n t i f y i n g whole groups r a t h e r than
individual items. It seems more p l a u s i b l e to consider 'change of
shape/size' verbs as a class a pa rt and a b o r d e r l i n e case o f motion
verbs. Others such as 'stand up' and ' s h a k e ' , which do not i nvolve
change-of-location of the whole object, are certainly within the
f i e l d o f motion verbs.

'Contact' verbs, a l re ady mentioned in connection w it h verbs

d e sc r ibi ng vari ous activities, are an even more probl emati c case as
they n e c e s s a r i l y imply motion (and us u a l l y change-of-location even)
o f at l e a s t one o f the o bj ect s i nv olved. Owing to t h i s implication
such verbs are included a t t h i s stage: they are f e l t to be nearer to
the ' c e n t r e ' o f the f i e l d than 'movement of boundaries' verbs or the
class comprising ' s m i l e ' , ' c ou gh ' , 'breathe', etc. Yet the m a j o r i t y
o f the verbs that belong to this group e x h i b i t quite idiosyncratic
characteristics, which do not seem to relate to ot he r properties
which pl ay a c e n t r a l role in e s t a b l i s h i n g inter-group relationships.
Hence no use w i l l be made of such verbs.

The c r i t e r i a discussed above can be employed t o c a t e g o r i z e MG verbs

of motion into CL and CP verbs. The m ajority of CL verbs imply
covering acertain distance. Th er ef or e in order to confirm the
correctness of a CL c a t e g o r i z a t i o n , one can consi der as a s u i t a b l e
syntactic test the c o m p a t i b i l i t y o f a particular verb w it h a phrase
de sc r ib i ng t h a t d i s t a n c e . One possi bl e t e s t -e n v i r on me n t i n v o lv i n g a
measure phrase which i s suggested in M i l l e r (1972) takes the form:

A _____ PAST (£) U a ) 10 p o n t u s / m e t r a / x i l j o m e t r a

A _____ -ed (B) (fo r ) 10 c e n t i m e t r e s / m e t r e s / k i l o m e t r e s .

- 76 -
Examples o f MG CL verbs which pass the above t e s t are 'tre x o ' (run),
'k ila o ' (ro ll), 's ik o n o ' (raise), etc. There are, however, verbs
which are i n t u i t i v e l y f e l t to descri be CL although they cannot occur
in such an environment. F ur the r syntactic tests are therefore
necessary to secure t h e i r inclusion. Four a p p r op ri at e environments
are proposed, c onc ent ra t ing on the poi nts where the motion starts
(source) and/ or ends (goal). List II contai ns all five test-
environments (a) t o (e) and the verbs which can occur i n each one o f
them. There is a n a tu r al hi e ra r ch y in these tests, in that most
verbs which are compatible wi th a certain environment are also
compatible wi th any subsequent one. The l a s t environment included in
List II, namely (e), allows also for the i n c lu s i on of CP verbs.
Examples of this last category are: 'talandevow e' (oscillate),
' s k iv o ' (bend), 'kaBome' (s it), e tc .^

In the present study only physical motion o f concrete obj ect s and a
literal understanding of MVs are taken into c o n si d e r a t i o n .
Discussing the boundaries o f ' l i t e r a l ' understanding l i e s outside the
scope o f the present a n a l y s i s , although the problem is i n f a c t very
acute in connection wi th MVs. Cases like (9) or ( 1 0 ) :

(9) o 5r om o s o b i v i s t o x o r j o
the road leads to the v i l l a g e

(10) o kisos s k a rfa lo s e mexri to p araB iro

the i vy (has) climbed up t o the window

i n vol ve concrete obj ect s and (10) describes also physical motion
( r e l a t e d t o gr owi ng). Yet the verbs 'o b iy o ' ( l e a d ) and' s k a r a f a l o n o '
( cl i mb) are not i n d i s p u t a b l y ' l i t e r a l l y understood' in such contexts.
Such cases w i l l be mentioned separately when the a nal ys is depends
crucially on whether a p a r t i c u l a r use i s literal or not; the whole
issue w i l l be discussed i n Chapter 3.

- 77 -
 Notes on Chapter 1

1. The terms 'componential a n a l y s i s ' and ' l e x i c a l decomposition' are

used i nt er cha nge abl y in t h i s t h e s i s ,

2. Some linguists have also used a mi xt ur e of two different

languages in their representations. So, Talmy (1972) uses the
Spanish preposition POR i n combination with English words to
r epr es ent what he calls "the deep morpheme" corresponding to
" t h r o ug h ".

3. M iller and Johnson-Laird (1976) use logical formulae in their

r ep r e s e n t a t i o n s w it h a b st r a c t operators introduced by 'if'
clauses (t hey discuss, however, the r e l a t i o n of these operators
to p a r t i c u l a r senses of the corresponding English words). From
this p oi n t o f view t h e i r semantic elements are a cross between
traditional features and condi ti ons (unspecified as to their
nature), I feel, th erefore, f r e e to r e f e r t o them using e i t h e r
term,

4. I am r e f e r r i n g to Frege's well known and frequently quoted

passage on ' l o g i c a l l y simple' elements:

"One cannot r e q u i r e t h a t eve ry thi ng s h a l l be decomposed any

more than one can r e q u i r e t h a t a chemist s h al l decompose
every substance. What is simple cannot be decomposed and
what i s l o g i c a l l y simple cannot have a proper d e f i n i t i o n .
Now something l o g i c a l l y simple is no more given us at the
outs et than most o f the chemical elements are; i t is reached
only by means o f s c i e n t i f i c work".
(Geach and Black 1 9 6 6 :4 3) .

5. It is quite p os si bl e that if these gods were conceived of as

having the 'su p e rficia l' characteristics of cows for i n st an ce ,
t h e i r o f f s p r i n g would not be c a l l e d 'b o y s '/'g irls '. Not ic e t h a t
the Mi nota ur , which lacked a number of the 's u p e rficia l'
c h a r a c t e r i s t i c s o f human beings, i s never c a l l e d ' b oy ' (as f a r as

- 78 -
 I know), although he shared w it h humans the p ro p e r t y of being
mortal d e s p i t e being the o f f s p r i n g o f a god.

6. Whether word meanings should be equated with necessary and

sufficient c ondi ti ons i s o f course an ext remel y i mportant issue
which deserves (and will be gi ven) s peci al c o n si d e ra t i o n in
subsequent s ec t i o n s .

7. I must admit t h a t I am one o f those few people who b e l i e v e t he re

is some value in this distinction, although I agree that the
attempts at defining the nature of either markers or
d i s t i n g u i s h e r s have never been s a t i s f a c t o r y .

8. In structuralist theories, gr adat ion is usually restricted to

specific types of vocabul ary, e.g . the ' h o t -w ar m -c o l d' kind o f
paradigm.

9. Lacking a n a t i v e speaker' s i n t u i t i o n s on the semantics of ' c l i m b '

I w ill consi der the data o f these analyses c o r r e c t and reserve
s tr onger claims f o r my MGMVs, f o r which I also have supporting
evidence from the semantic s i m i l a r i t y and p r o t o t y p i c a l i t y t e s t s I
have c a r r i e d out.

10. E x t e n s i o n a l i s t t h e o r i e s of meaning ( e . g . Putnam 1970, 1975 Kripke

1972) are often juxtaposed to 'm entalistic' ones (Dahlgren
1978:58). This term w i l l not be used here as it might c re at e
confusion regar di ng the ultimate goal of lexical linguistic
investigation which is undoubtedly that of characterizing the
'mental' lexicon. I can see no discrepancy between ' m e n t a l i s t i c '
and 'non-mentalistic' theories o f meaning concerning t h i s goal,
i.e. arriving at a description of semantic competence: of
whatever i s in the mind o f an i n d i v i d u a l which enables him/her t o
use a p a r t i c u l a r word a p p r o p r i a t e l y .

11. I am d e l i b e r a t e l y r e s t r i c t i n g t h i s short di scussion o f the theory

o f s te reot ypes to Putnam ( 1 9 7 5 ) , as most poi nts of interest to

- 79 -
 the present investigation (also analysed in Putnam 1970 and
Kripke 1972) are covered t h e r e .

12. This p oi n t needs c l a r i f i c a t i o n . To s t a r t w i t h , what Putnam says

in t h i s connection i s not s t r i c t l y t r u e w i t h i n the framework o f
his own system: specialists could wel l possess what c o n s t i t u t e s
necessary and sufficient c ondi ti ons for a term fallin g within
their domain, since they also have i n fo rm at io n fixing the
extension of the term. So Putnam's c la i m that knowledge
s u f f i c i e n t to fix the extension is not possessed by a l l or any
speakers does not apply to a l l terms. What he i s probably r i g h t
about is t h a t the 'psychol ogi cal state' of the i n d i v i d u a l speaker
does not determine the ext ens ion. I t h e r e f o r e tend to understand
t h i s r a t h e r sweeping statement as an attempt a t drawing a t t e n t i o n
to the f a c t t h a t meaning i s not a fu nc t io n o f the psychological
s t a t e o f any i n d i v i d u a l speaker, in ot he r words t h a t knowing the
meaning of a term is not a ma tt er of being in a certain
psychological state. Another i m p l i c a t i o n , however, is completely
c o n si s t e n t w i t h hi s t he o ry , although not wel l i l l u s t r a t e d by the
terms just mentioned. Putnam claims that the idealization of
supposing t h a t there is such a t hi n g f o r which the d e s c r i p t i o n
'tre e ' is t r u e (i.e. the p r op os i t i on ' Thi s is a tre e ') is very
severe. This cannot, however, be e q u a l l y the case f o r terms such
as 'gold' or 'water' which have an indisputably identifiable
extension ( a cc e s s i b l e to s c i e n t i s t s ) a t l e a s t concerning ' c u r r e n t
s c i e n t i f i c knowledge'.

13. I am obvi ousl y t a l k i n g about the l ex ic on o f the same language.

Besides this, the l e xi con of each language w ill necessarily
reflect in large part both culture and language specific
i nf or ma ti on (see also Pulman 1 983:168).

14. I am using 'weak' here in Putnam's sense, where r oughly speaking,

the s tr onge r t he s t er eot ype the near er one is to actually
possessing necessary and s u f f i c i e n t c o n d i t i o n s , o r , a t any r a t e ,
enough i n f or m at io n to d i s t i n g u i s h the item i n question from a l l
r e l a t e d ones.

- 80 -
15. A 'category' is understood as a number of obj ect s which are
considered equivalent and designated by a name, e.g. 'c a t',
'c h a ir'.

16. The actual passage reads as f ol l o w s:

" . . . words in na tur al language are not g e n e r a l l y ' y e s - n o ' :

t h e r e are t hi ngs o f which the d e s c r i p t i o n ' t r e e ' i s c l e a r l y
f a l s e , to be sure, but t h er e are a host o f b o r d e r l i n e cases.
Worse, the l i n e between the c l e a r cases and the b o r d e r l i n e
cases i s i t s e l f f u z z y . Thus the i d e a l i z a t i o n i nvolved in
supposing t h a t t h e re i s such a t h i ng as the set o f things o f
which the term ' t r e e ' i s t r u e - i s a c t u a l l y very severe".
(Putnam 1975b:133)

17. By t h i s term I am r e f e r r i n g to a characteristic of the basic

level a l re ady mentioned, namely t h a t its c at e go r ie s contain the
most i n f o r m a t i o n .

18. For the c o l l e c t i o n o f m a t e r i a l , standard d i c t i o n a r i e s o f MG have

been used: Bostanjoglou (1962) and Dimitrakos (1969). For
special purposes, e.g. distinguishing between basic level and
subordinate terms, the ' cor pus' was supplemented wi th ma t er ia l
from t h r e e fiction books: Tzortzogl ou N. "Otari O rgizete i G i"

(When the Earth Gets Angry), Kazantzaki N. "Megas Aleksandros"

(Alexander the G r e a t ) , Sari Z, "otan o ilio s ..." (when the sun
...). These were chosen simply because they contai n a l o t of MVs
in ext ens ive descriptions. Three newspapers ' Vim a',
' K a t h i m e r in i ', and 'Ethnos' were searched for supplementary
material from September to December 1983 and two magazines
'Ik o n e s ' and ' G i n e k a ' in 1985, to secure i n c lu s i on o f more recent
uses. My personal intuitions concerning this material were
c on st a nt l y checked agai nst those of ot he r native speakers of
standard MG. It is therefore assumed t h a t the sample o f MGMVs
used in the present a n al y si s is a t l e a s t r e p r e s e n t a t i v e .

- 81 -
 2. THE 'STATES - PROCESSES - EVENTS' CATEGORIZATION
AND ITS APPLICATION TO MOTION VERBS

2.1 Preliminaries

It is generally assumed and sometimes e xp lic itly stated (Lyons

1 97 7:4 82 -4 ) t h a t our pe rcepti ons of the physical world are organized
and described w i t h i n a conceptual framework which provides for the
identification of (a) states-of-affairs and (b) processes, events,
actions. Hence the 'archetypal' distinction between static and
dynamic s i t u a t i o n s is almost unanimously accepted in most attempts a t
a semantic c l a s s i f i c a t i o n of verbs. However, this is about a l l that
there is s ig n ific a n t agreement on. As w i l l be seen in what f o l l o w s ,
there is no agreement on what 'dynamic' i nc lude s, nor on what the
distinction is a distinction of. Sentences, predications and verbs
are the usual candidates but also situations at large. Somewhat
different positions are adopted by Vendler (1957), Kenny (1963),
S e ile r (1968), Leech ( 1 9 6 9 ) , Coseriu ( 1 9 7 1 ) , M i l l e r and Johnson-Laird
(1976), Lyons ( 1 9 7 7 ) , Mourelatos ( 1 9 78 ) . No s i n g l e approach, though,
can c o n s t i t u t e an adequate basis f o r the d e s c r i p t i o n of the f a c t s o f
MGMVs, so the main poi nts o f controversy have to be looked a t in some
d e ta il.

It is i mportant to investigate the e xt e n t to which the d i s t i n c t i o n

bears on the semantics o f i n d i v i d u a l Vs or c h a r a c t e r i z e s verb- for ms,
or grammatical categories, since t he re is some t r u t h in different
cla ims , e . g . w it h regard t o a ss oc iat ions w it h ' a s p e c t ' , wi th the verb
versus a d j e c t i v e d i s t i n c t i o n as well as w i t h the semantic f e a t u r e s o f
individual Vs (duration vs instantaneity of action). Besides the
i n t e r a c t i o n o f these f a c t o r s , additional ones can be det ect ed in the
area o f MGMVs, such as the r o l e o f the natur e o f the 'theme' (moving
object).

In English the d i s t i n c t i o n is assumed to l e x i c a l i z e in the opposi ti on

between ' b e ' and ' h a v e ' , on the one hand, and 'become' and ' g e t ' , on

- 82 -
the other. In the domain under investigation one can consider MG
'position' Vs such as: '/me' ( b e ) / 'v ris k o m e ' (be found, located),
's te k o '/'s te k o m e ' (stand), 'k a B o m e j'/im e kaBism enos-i ' (be s ea t ed) ,
'ye rn o /im e yerm enos-i' (lean), 'k s a p lo n o j/im ksaplom enos-i' (lie) as
candidates for the lexicalization of stativity and all MGMVs as
dynamic verbs, i.e. candidates f o r the l e x i c a l i z a t i o n o f the absence
of s t a t i v i t y . Everything e ls e i s c o n t r o v e r s i a l .

The major issues t h a t w i l l h o p e f u l l y be i l l u m i n a t e d are:

1. Whether the distinction into two c a t eg or i es ('states' and

e ve r yt hi ng e l s e ) is adequate.

2. Whether the boundaries between the c at e go ri e s are clear-cut or

f uz zy and the distinction a ma tter of gr adat io n rather than a
categorical 'yes-no' phenomenon.

3. What i s the ' o b j e c t ' o f the d i s t i n c t i o n . (Does i t a f f e c t verbs,

p r e d i ca t i o n s or both?)

4. Which p a r t i c u l a r characteristics o f verbs or whole p r e d i ca t i o n s

contribute to the distinction - semantic characteristics and
aspect of individual verb forms, presence and type of PP,
semantic characteristics of the theme (such as 'countabi.lity')
being the most obvious c andi dates.

2 .2 An overview o f the standard t e s t s f o r d i s t i n g u i s h i n g between

s t a t e s - processes - events and how t o f a i l them

Before t a c k l i n g the problems 1 to 4 ( s e ct i o n 2 . 1 ) it i s i mportant to

draw attention to two points. Firstly, the distinction between
s t a t e s and events is u s u a l l y understood as an i n t u i t i v e one, rarely
de fi ned and more o f t e n e x e m p l i f i e d by means o f sentences o f the type:

- 83 -
 M a r i e t t a understands Quantum Mechanics ( s t a t e )

vs Marietta arrived at 5 o'clock (event).

Secondly, a number o f s y n t a c t i c t e s t s have been used by a number of

l i n g u i s t s to help i d e n t i f y s t a t i v e and n o n - s t a t i v e verbs, a t l e a s t in
E ngl ish. A c o n si d er a ti o n o f these t e s t s seems in order here, as i t
has a bearing on a l l f o ur problems.

The standard t e s t f o r i d e n t i f y i n g s t a t i v e Vs in English is to apply

the r u l e f o r c r e a t i n g pr ogressi ve forms: s t a t i v e s normall y r e j e c t the
pr ogressi ve aspect, hence t h i s method i s e x t e n s i v e l y a ppl ied (Kenny
1963, M iller and Johnson-Laird 1976, among others). As the
progressi ve aspect covers (among ot he r t h i n g s ) the need to r epresent
a s i t u a t i o n as happening or developing through t i m e, i t is n a tu ra lly
i ncompatible with 'state' denoting verbs. Anderson (1973), relates
const ruct io ns exhibiting pr ogressi ve aspect to 'p a ra lle l' ones
i n v o l v i n g "be in the process of Ving", e . g . :

John is fallin g - John i s in the process of f a l l i n g

*John i s knowing the t r u t h - *John i s in the process o f knowing

the t r u t h .

This t e s t could be used as a na tur al test for singling out English

sta tives or, at least, for d i s t i n g u i s h i n g between s ta t es and
eve ry thi ng e l s e . N ot i c e , however, that problems a r i s e a lr ea dy: I 'm
understanding more about Quantum Mechanics as each day goes by is
offered by Mourelatos (1978:419) as an example of "semantic
transposition", which causes ' understand' t o f u n c t i o n in an a c t i v i t y
context u n l i k e He understands the r e a l natur e o f t he problem which i s
given as an i nstance of a 's ta te '. Yet even Mourel atos, who
recognizes that we are actually d e al i ng with a categorization of
predications rather than words, and moreover that aspectual

- 84 -
differences are a l l i mpor tant, accepts t h a t the speci al affin ity of
'understand' or 'know' w it h s t a t e s is beyond doubt.

M iller and Johnson-Laird (1976) suggest that the s t a t iv e- dy n am i c

distinction might be 'prim itive' and t h a t although its linguistic
representation by means o f s t a t i v e vs n o n - s t a t i v e verbs may not be
universal, i t i s present in English verbs a t l e a s t . T h e i r suggestion
seems to stem from Brown's (1973) observations to the effect that
both the pr ogr essi ve '-in g ' and the I mperati ve are present in the
earliest sentences children utter, but never wi th stative verbs,
although a number of ot he r const ruct ions are ov er ge n er al i ze d and
therefore used w i t h u n s u it a bl e stems. Not ice, therefore, that the
validity o f the pr ogr essi ve and Imperati ve t e s t s is not questioned.
The occurrence o f verbs wi th the progressive and the I mperati ve is
not even understood as a simple test but rather as the d e ci s i v e
f a c t o r on the basis o f which an i n t u i t i v e d i s t i n c t i o n can acquire a
'form al' status. Thus ??Know the answer and ??Be seven f e e t t a l l are
offered by M i l l e r and Johnson-Laird (1976:474) as examples of the
incompatibility between sta tives and the I m p er a t i v e. It is not,
however, immediately obvious t h a t such i s o l a t e d 'form al' means w i l l
n e ce s s a r i l y y i e l d unquestionable r e s u l t s . N ot i c e , f o r i nst an ce , t h a t
although Know thyself is not what textbooks call 'good Modern
English', is by no means as obviously unacceptable as the previous
two examples. S i m i l a r l y Be t he re at seven o ' c l o c k sharp seems q ui t e
acceptabl e. In s h or t , incompatibility with the I mper ati ve w ill
obviously depend on whether the addressee is in a p o s i t i o n to car ry
out the order (issued through the Imper ati ve) in the firs t place.
Hence the u n a c c e p t a b i l i t y o f Be seven f e e t t a l l provides no argument
concerning stativity in general and the semantics of ' be ' in
p a r t i c u l a r as M i l l e r and Johnson-Laird intend it. It can obviously
be argued t h a t we are d e al i ng w i t h two d i f f e r e n t 'be ' verbs or with a
'positional' vs a 'non-positional' sense of 'b e '. The point is,
however, that the ambiguity is clarified on the basis of the
p r e d i c a t i o n as a whole and t h a t the I mperati ve t e s t i s na t u r a l only
if it checks things, l i k e ' i n t e n t i o n a l i t y ' , the p o s s i b i l i t y o f a verb
to d i s p l a y a ' v o l u n t a r y ' vs a ' n o n - v o l u n t a r y ' characteristic, etc.*
Not ice also t h a t the pr ogr essi ve t e s t i s not much s a f e r . Verbs such
as ' f i n d ' (used n o n - i t e r a t i v e l y ) do not take the pr ogressi ve any more

- 85 -
than stative verbs, although they are intu itively understood as
' e v e n t ' r a t h e r than s t a t i v e .

The situation is even less promising when one turns to MG Vs o f

position and motion. The progressi ve t e s t i s obvi ousl y i n a p p l i c a b l e
i n the case o f MG as i t does not have t h i s p a r t i c u l a r formal means o f
r e p r es e nt i ng situations developing through t im e. Besides, the
'p a ra lle l' method o f e x p l i c i t l y using phrases l i k e 'in the process
o f', 'in a state of' (which has been proposed for Engl ish) would
yield quite bizarre MG sentences which should not be considered by
the s e m a n t i c i s t (as al re ady argued in Chapter 1 ) . Rather e xpectedl y,
the I mpe ra t iv e t e s t stands no chance as a formal criterion; consider
the case of 'k s e ro ' (know) which is probably the stative 'par
excellence': 'ksere to pos 6a f i y i ' (know i t , t h a t s/he w i l l l eave)
or the equivalent 'na kse ris pos 8a f i y i 1 are perfectly good
sentences. Sim ilarly ' na i s e s i n e x o s k s a p l o m e n o s j a na k a l i t e r e p s i s 1
(be conti nuousl y l y i n g ( i n bed) t o get b e t t e r ) is f i n e , although '/me
ksaplom enos' looks l i k e a s t a t i v e p r e d i c a t i o n to me.

It would seem, at firs t sight, that tests based on the observation

that, unlike events, s t a t es do not 'happen' would yield safer
results. Hence M i l l e r and Johnson-Laird ( i b i d . ) use a t e s t question
of the form: What happened? predicting that 's ta te ' p r e d i ca t i o ns
cannot q u a l i f y as answers. T h e i r example i s : ??He be li ev ed in Santa
Claus. Yet the na tur e of such t e s t s (Question and Answer) i s such
t h a t one would r e q u i r e a p r i o r a nal ys is of c onver sat ional principles
a t work in each p a r t i c u l a r case, before using them as s a f e, formal
diagnostics. Not ice t h a t He knew eve ry thi ng by h e ar t does not need a
very e l a b o r a t e supporting scenario to pass as a p e r f e c t l y acceptable
answer t o What happened? although 'know' and ' b e l i e v e ' are 'model'
statives. Possi bl e counter-arguments t o the e f f e c t t h a t the 'real'
answer to the questi on is in fact something about the logical
s u b j e c t ' s performance, f o r i nst an ce , which i s a c t u a l l y i mpl ied by He
knew e ve r yt hi ng by he ar t are beside the p o i n t ; the f a c t remains t h a t
'purely lin g u is tic ', 'fo rm al' and ' v e r b - c e n t r e d ' tests a re, f o r the
most p a r t , i n a p p r o p r i a t e or , at best , misconceived as h e u r i s t i c s f o r
the ' s t a t i v e - d y n a m i c ' d i s t i n c t i o n . So, q u i t e e xpectedl y: 'ti s in e v i/
ti e yin e ? ' (what happened?) can t r i g g e r a number o f answers w i t h what

- 86 -
seem to me s t a t i v e p r e d i c a t i o n s , e.g. ' k a Q o t a n / k a Q e t e / i ne k a Q i s me n o s

meres o lo k lire s a k in ito s se mia k a re k la ' (he was seated/sits/is

seated f o r days on end motionless on a c h a i r ) . Apparentl y 'happen'
is so general that even if it is used in conj unct ion w it h the
predication whose s ta tus we seek to identify, the ambiguity is not
solved through formal means. Hence He happens to understand the re al
nature o f t hi ngs i s p a r a l l e l to I t so happens t h a t he understands the
r ea l nature o f t hi ngs and even i f one is not too happy about e i t h e r
sentence, one only has t o consi der He happens t o be seven f e e t t a l l
to r e a l i z e t h a t even in English we are back where we s t a r t e d . 2

I have discussed the standard t e s t I happen to be aware o f in some

detail in the hope o f showing:

( a) That simple or i s o l a t e d 'formal' c r i t e r i a are not as adequate as

they are sometimes expected t o be even w it h respect to the major
d i s t i n c t i o n between s t a t i v e s and n o n - s t a t i v e s .

(b) That the problem cannot be adequately approached through

c once ntr ati on on verb-forms alone. M iller and Johnson-Laird' s
remark t h a t although, s t r i c t l y speaking, it i s the p r e d i c a t e as a
whole t h a t is s t a t i v e or n o n - s t a t i v e " i t i s t edi ous t o keep such
subtleties always in mind" ( p . 475) seems r a t h e r unwarranted in
t h i s p a r t i c u l a r case.

It seems more likely that the stative/non-stative distinction

affects verbs through p r e d i c a t i o n s ; hence a number o f d i f f e r e n t
factors have to be taken into c ons id e ra t io n (such as those
mentioned in 4 o f the previous s e c t i o n ) . I t also seems important
t o l ook f o r acceptabl e attempts at defining the content o f the
distinction(s) in question and then try to 'p a ir' definitions
wi th formal syntactic c rite ria (e .g. what the verb is in
c onst r uct io n w i t h ) .

- 87 -
2.3 F ur the r c r i t e r i a f o r the d e f i n i t i o n o f s t a t es- pr oce sse s- eve nt s

A number of linguists, e.g . Coseriu (1971), Leech (1969), Seiler

(1968) and Langacker (1975), operate with a dichotomous
categorization of predications into 'states-events' or 'states-
proc es ses '. In what f o ll o ws i t w i l l become obvious t h a t we need more
than a two-way c l a s s i f i c a t i o n . A f a i r l y concise account o f what the
notions 's ta te ', 'process', 'event' i nv ol ve is offered in Lyons
(1977) and can be used as a firs t step towards defining these
noti ons . Lyons accepts the major distinction between static and
dynamic s i t u a t i o n s and then proceeds t o s p ec if y the content of the
di s t i n c t i o n :

”A s t a t i c s i t u a t i o n ( or s t a t e - o f - a f f a i r s , or s t a t e ) is one t h a t
i s conceived o f as e x i s t i n g , r a t h e r than happening, and as being
homogeneous, continuous and unchanging throughout i t s d u r a t i o n .
A dynamic s i t u a t i o n , on the ot he r hand, i s something t h a t happens
(or occurs, or takes p l a c e ) : i t may be momentary or enduring; i t
i s not n e c e s s a r i l y homogeneous or continuous, but may have any o f
several temporal contours; and, most i mportant o f a l l , i t may or
may not be under the contr ol o f an agent. I f a dynamic s i t u a t i o n
is extended in t ime, i t is a process; i f i t is momentary, i t is
an event; and, i f i t is under the c ontrol o f an agent, i t is an
action. F i n a l l y , a process t h a t i s under the contr ol of an agent
i s an a c t i v i t y ; and an event t h a t i s under the control of an
agent i s an a c t . " ( p . 483)

This d e f i n i t i o n j uxtaposes 'existing' to 'happening' and 'momentary'

to ' e n d u r i n g ' . These opposi ti ons can be used as a working hypothesis
to differentiate between sta te s, processes and events. The
additional notions of 'c o ntrol', 'agent', 'act' and 'a c tiv ity ' are
only important in a discussion o f what I consider a separate i ssue,
namely agentivity, and are dealt w it h in detail in the f ol l o w i n g
c hapt er . As I have a lr ea dy r e f e r r e d to 'a c tiv itie s ', however, (in
the understanding that the term i s more w i de ly used and e a s i e r to
i n t e r p r e t on the basis o f common everyday use) it seems necessary to
show s ch e ma t i ca l l y where 'a ctivity' stands in relation to Lyons'
remaining c a t e g o r i e s .

- 88 -
 Process + agent = a c t i v i t y
Event + agent - act
A c t i v i t i e s and acts are actions

Having i s o l a t e d the factor of 'co n tro l', we can c oncentr ate on the
content o f the t r i p a r t i t e d i v i s i o n i n t o s t a t e s , processes and events.
The only f e a t u r e which d i s t i n g u i s h e s s t a t es from e ve r yt hi n g e ls e is
that they cannot be said to 'happen' or 'occur'. Processes are
distinguished from events as being enduring rather than momentary.
Events and processes can be said to mark a change o f s t a t e , i.e. a
t r a n s i t i o n from an i n i t i a l s t a t e to a f i n a l one. If this transition
has some d u ra t i on i t is a process; i f not, an event.

It is r a t h e r obvious t h a t in order to categorize predications (and

ultimately specific verbs) we need f u r t h e r elaboration o f what the
relevant notions involve. Philosophers o f language such as Vendler
(1957), and Mourelatos (1978) have discussed these d i s t i n c t i o n s in a
r a t h e r i l l u m i n a t i n g way. The main points o f these accounts w i l l be
looked a t b r i e f l y and connected t o the f a c t s o f MGMVs.

Vendler (1957) observes that while ' r u nn i ng ' has no set termi nal
point, ' runni ng a mile' has a ' c l i m a x ' which has to be reached if
"the a ct io n is to be what i t is claimed to be" ( p . 145); i.e. running
a mi l e has t o be f i n i s h e d to be t r u e . N e i t h e r V e n d l e r ' s terminology
nor his c a t e g o r i z a t i o n w i l l be adopted here, but it is i mportant to
n o t i c e f o r a s t a r t t h a t the d i s t i n c t i o n between a changing s i t u a t i o n
(runni ng) and a s i m i l a r one i n v o l v i n g a c u l m i n a t i o n , e.g . running a
mi le (or an end p oi n t presumably, e.g. running to the kiosque) is
recognized for what it is; the former is not a rb itra rily named a
'state' and ' r u n ' does not r i s k acqui ri ng t he f e a t u r e [+stative] as
in S e i l e r ( 1 9 6 8 ) . The immediate i m p l i c a t i o n f o r a sentence l i k e (1):

(1) ta p e b ja trexun ston kipo

the c h i l d r e n r u n / a r e running t o / i n the garden

is th a t it cannot be b a s i c a l l y considered as i n v o l v i n g a state. If

it i s understood as i n v o l v i n g a goa l, it can be i n t e r p r e t e d as being

- 89 -
a kind o f event (in Vendler's terms an 'accompl i sh me nt ' ); if it is
understood as i n v o l v i n g a l o c a t i v e , i t can be i n t e r p r e t e d as being a
kind o f process ( i n V en d l e r ' s terms an ' a c t i v i t y ' ) . Vendler is not
responsi bl e f o r these interpretations, but I consi der t h a t they are
in keeping w it h hi s (1957) account of what processes ar e, i.e.
homogeneous in t h a t t h e i r time s t r e t c h is inherently indefinite and
in t h a t t h e r e is no c ul mi na ti on and no a n t i c i p a t e d r e s u l t . We w i l l
r e t u r n t o t he i n t e r p r e t a t i o n of examples such as (1) a f t e r completing
the p r e se n ta t i on o f the r e l e v a n t noti ons.

A further distinction is drawn between ' accomplishments', which

i n volve time per iods, and 'achievements' which i nv ol ve time i n s ta n ts
(e.g. reaching the h i l l t o p ) . The i m p l i c a t i o n for sentences (2) and
(3):

(2) a ne vike s tin k o rifi mesa se m i a o r a

he went up t o the top w i t h i n one hour

(3) e ftase s tin k o rifi mesa se m i a o r a

he reached the top w i t h i n one hour

i s t h a t although they are s i m i l a r in many r es pe c ts , the former one is

an 'accomplishment' and the latter an ' a c h ie v em e nt ' . While (2)
i mpl ies t h a t whatever i s i nvolved in 'aneveno' (go up) took one hour
(a time p er io d) to complete, i.e. that ' goi ng up' went on f o r an
hour, (3) does not imply t h a t whatever is i nvolved in 'fta n o ' (reach)
went on f o r one hour; 'reaching' takes a time i n s t a n t (i.e . a poi nt
and not an i n t e r v a l ) t o complete.

The d i f f e r e n c e between (2) and (3) can u l t i m a t e l y be a t t r i b u t e d to

the i nh e re nt semantic features of the two verbs i nvolved and in
particular to some feature 'momentary' vs 'enduring'. Not ic e,
however, that d ur a t i o n (by itself) is not sufficient in
d i s t i n g u i s h i n g between events and processes, as i t appears to be in
Lyons ( 1 9 7 7 ) . It i s used here t o d i s t i n g u i s h between 'achievements'
and 'accomplishments' which are both subsumed under events by Kenny
(1963) and Mourelatos (1978). Mourelatos not ic es that both

- 90 -
c at egor i es in question ('developments' and 'punctual oc cur rences' ,
respectively, in hi s ter mi nol ogy) take d e f i n i t e time but w hi le the
former are i n t r i n s i c a l l y characterized by d u r a t i o n and r e f e r to the
whole o f the time segment, the l a t t e r cannot occur ov er /thr oughout a
temporal stretch, but only at a s i n g le moment (w ithin a temporal
stretch).

Hence the main p oi nt s a t which events d i f f e r from processes are: (a)

homogeneity (which everybody seems to accept as characterizing
processes o n l y ) , i.e. the p r e r e q u i s i t e t h a t any p a r t o f a process be
of the same nature as the whole and (b) that the time stretch of
processes ('a c tiv itie s ' in Vendler) is indefinite. Contrary to
these, events i nv ol ve some c ul mi nati on or an a n t i c i p a t e d r e s u l t and
definite t ime. States, fin a lly , are generally understood as not
constituting a change, although “they may a r i s e or be acquired as a
result of change" (Mourelatos 1 97 8: 416 ). Clearly they may also
endure or p e r s i s t over s t r e t ch e s o f ti me.

Adopting Mourel atos' t er mi nol ogy, where processes and events are
understood as 'occurrences', c o n tr a ry to s ta t es which do not occur/
happen ( c f . Lyons 1977) and developments and punctual occurrences are
subcategories of events, we can present the relevant c riteria
s ch e ma t ic al l y as f o l l o w s :

CHANGE/ HOMOGENEOUS INDEFINITE DURATION

OCCUR TIME

States + + +
Processes + + + +
rDevelopments + +
k Lpunctual Occurrences +
note: a= occurrences,
b= events

The f a c t t h a t s t a t es do not c o n s t i t u t e a change i s understood here as

saying l i t t l e more than t h a t they do not happen/occur, hence a s in gl e
dimension is used in the above schematic representation. Besides,

- 91 -
the f a c t t h a t events have d e f i n i t e times i s understood here as l i n k e d
t o the p r e r e q u i s i t e t h a t t he re be no c ul mi na ti on and no a n t i c i p a t e d
result, hence t he latte r properties are not allocate d a separate
column. So if we wish to differentiate between 'running' and
'running-a-mile' (Vendler's examples of an activ ity and an
accomplishment r e s p e c t i v e l y ) homogeneity by i t s e l f is not adequate.
As Mourelatos observes ( f o r d i f f e r e n t purposes):

"the g e n er i c a c t i v i t y o f running can be f u r t h e r d i f f e r e n t i a t e d

i n t o a species (one among i n d e f i n i t e l y many) o f r u n n i n g - a - m i l e
w i thout i t s l o s i n g i t s c h a r a c t e r as an a c t i v i t y . In ot he r words,
r egar dl ess o f whether a mi l e i s or f a i l s t o be run, any
subst ret ch o f r u n n i n g - a - m i l e a c t i v i t y d i v i d e s homogeneously i n t o
sub- st re tc hes o f the same" ( i b i d . : 420)*

The r e s t r i c t i o n imposed on the activ ity of running by 'a mile' is

comparable to that of a goal specification, e.g. running to the
kiosque. Unl i ke both, though, running i n the garden can be l i n k e d to
the unrestricted and open-ended a c t i v i t y of 'running' in t he sense
t h a t t h e r e i s no f or s e e a b l e and s ta t ed t e r m i n a t i o n .

With these s p e c i f i c a t i o n s in mind we can now exe mpl i fy the r e l e v a n t

notions with MGMV p r e d i c a t i o n s and present them s ch e ma t i ca l l y as
f ol l o w s:

States

( 4) kaQete s i n e x i a brosta s to paraQ iro

s/he s i t s c o n s t a n t l y i n f r o n t o f the window

- 92 -
 Occurrences

Processes Events

(5) t r e x i mesa s t o n k i p o developments punctual occurrences

s/he r u n s / i s running (6) ska rfa lo se s tin (7) v u t i k s e s t o p o t a m i
in ( w i t h i n ) the garden k o rifi tu vunu s/he dived i nt o
s/he climbed to the r i v e r
the top o f the
mountain

2>4 Adv er bi al s o f time and goal vs l o c a t i v e

In order t o see more c l e a r l y the i m p l i c a t i o n s o f examples (4)-(7) it

is useful to see what sorts of time a dve rbi al phrases they are
compatible with. Four kinds of such a d ve r b i al s need to be
d i s t i n g u i s h e d f o r the present purposes:

(a) those r e f e r r i n g to the du ra t io n f o r which a s i t u a t i o n is

valid, e.g. ' e p i mia o r a ' ( f o r one hour);

(b) those r e f e r r i n g to the frequency o f periods o f time during

which a s i t u a t i o n i s v a l i d ( u s u a l l y c a l l e d 'fr equency
a d v e r b i a l s 7) , e.g . ' ka Qe m e r a ' (everyday), 's ix n a ' (often);

(c) those r e f e r r i n g to a po i nt in time a t which a s i t u a t i o n i s

v a l i d , e.g . 's tis 6' (at 6);

(d) those r e f e r r i n g to the number o f times a s i t u a t i o n (which i s

r epeated) is v a l i d (usally called 7c ar d i n a l count
a d v e r b i a l s 7) , e.g. '4 f o r e s ' (4 t i m e s ) .

- 93 -
N o t i c e, firs t, that if (4) is to be understood as a stative
predication it cannot accept any o f the a d ve r b i a l s (a)-(d) s trictly
speaking. The moment ( a) or (b) type o f phrases r e pl a c e ' s i n e x i a '
the implications of an unlimited existing situation are removed.
( C l e a r l y (c) and (d) are compl etel y i n a p p l i c a b l e . )

The p a t t e r n which emerges i f ( a ) - ( d ) are attached to examples ( 5 ) - ( 7 )

looks l i k e t h i s :

(a) (b) (c) (d)

e p i mi a ora kaQe m er a s tis 6 4f o r e s
f o r one hour every day at 6 4times
Process
( 5) tre xi mesa s t o n k i p o + +
s/he r u n s / i s running
w i t h i n the garden

Event: Development
(6 ) s k a r f a l o s e s t i n k o r i f i - ? +
tu vunu
s/he climbed t o the top
o f the mountain

Event: Punctual Occurrence

(7) vu tikse sto potam i - + +
s/he dived i n t o the r i v e r

The combination of (6) wi th (c) results in (6a): 's k a rfa lo s e s tin

k o rifi tu v unu s t i s 6' ( s / h e climbed to the top o f t he mountain at 6)

which is not a completely acceptable sentence. If the goal
expression i s replaced w i t h a d i f f e r e n t one, e.g . 's to tra p e z i' (on
the ta b le ) there i s no problem w it h a s p e c i f i c a t i o n o f the poi nt in
time the event took place f o r obvious common sense reasons: the event
can be understood as a punctual occurrence and w i l l behave e x a c t l y
like (7). Hence (6) is u s e d . as a more or l ess clear case of a
development. As a lr e ad y poi nted out, however, if a ti me period
a dve rbi al such as ' mesa se mia ora' ( w i t h i n one hour) is used, the

- 94 -
 d i f f e r e n c e between sentences such as ( 6) and (7) is brought out most
clearly:

(6b) ska rfa lose s tin k o rifi tu v u n u mesa se m i a o r a

s/he climbed to the top of the mountain w i t h i n one hour

(7 a) vu tikse s t o p o t a m i mesa se m i a o r a

s/he dived i n t o the r i v e r w i t h i n one hour.

In these r a t h e r c a r e f u l l y chosen and o v e r s i m p l i f i e d cases (examples

( 4) to ( 7 ) ) d i f f e r e n t verbs have been used in the hope o f g e t t i n g as
' a c c u r a t e 7 an i l l u s t r a t i o n of the main points o f the d i s t i n c t i o n as
po s si b le . Not ice how d r a m a t i c a l l y the overall picture changes if
d i f f e r e n t verb forms of the same verbs are used:

(a) (b) (c) (d)

e p i m ia o ra kaQe m er a s tis 6 4 fores
f o r one hour every day at 6 4 times

(8) e t r e k s e mesa s t o n k i p o

s/he ran w i t h i n the garden

(9) ska rfa lon e s tin k o rifi ?? + ??

tu vu nu

s/he climbed/was climbing

to the top o f the mountain

Two main t hi ngs have to be not ic ed: firs tly, that the new f a c t o r
brought i n t o pl ay i s 'aspect'. The d i f f e r e n c e between ( 6) and (9) is
an aspectual one; the time remains constant, i . e . pa st . A brief
p r es en ta t io n of the main f a c t s o f MG Vs ( i n this respect) w ill be
given shortly, focusing on the few poi nts necessary for an
a p p r ec i a t i o n o f the distinctions under c o n s i d e r a t i o n . Secondly, it
must be noti ced that we have a lr ea dy moved into an area where
situations ( c o n t r a r y to the c a r e f u l l y chosen example sentences (4)-

- 95 -
(7)) are much less easy to i d e n t i f y and consequently match a c c ur a t e ly
w ith the 'pr oc ess ' - 'event' labels.

A combination of (8) with ( a) may be a cc eptabl e. If a goal

expression is, however, substituted f o r the l o c a t i v e , the r e s u l t i n g
sequence is compl etel y unacceptable, i.e.:

( 10) e trekse sto p e rip te ro e p i mia o ra

s/he ran t o the kiosque f o r one hour.

This is a further indication that goal constitutes an additional

obstacl e to the possibility of a situation being understood as a
process; the predication 'e tre kse sto p e rip te ro ' is completely
incompatibl e wi th the noti on of duration. For the same reason, a
combination of 'e tre kse sto p e rip te ro ' w it h (c) (i.e . a poi nt in
t ime) i s q u i t e acc epta bl e. This new s i t u a t i o n i s probably understood
as a fa irly clear case of an event, as w ill be shown after a
di scussion o f the aspectual change also i nv olved.

A combination o f (9) w it h (a) is probl emat ic only owing t o the f a c t o r

o f goal . I f a locative, e.g. 'to vuno' (t he mountain) i s s u b st i t u t e d
for i t , the r e s u l t i n g sentence is p e r f e c t l y acceptabl e:

(11) ska rfa lon e to vu no e p i m i a o r a

s/he climbed/was cli mbi ng (up) the mountain f o r one hour.

Once the obstacle/restriction imposed by the specification of the

goal is removed, 's k a rfa lo n e ' is ' n a t u r a l l y ' combinable wi th dur ati on
and the possibilities of understanding the vari ous resulting
p r e d i c a t i o n s as processes are i ncreased.

As a f i n a l remark on examples ( 8) and (9) n o t i c e t h a t a combination

of (9) with ( c) needs an e l a b o r a t e scenario to interpret even f o r
those speakers f o r whom i t i s m a r g i n a l l y a cc e pt a bl e. I t appears t h a t
the combination o f a goal r e s t r i c t i o n wi th a p o i n t in ti me adverbi al

- 96 -
is a strong factor of 'eventualization' and clashes wi th the
'pr oc ess ' i m p l i c a t i o n o f the verb.

An exhausti ve analysis of the implications of all possi bl e

combinations o f examples ( 5) - (10) would r e q u i r e a previous d e t a i l e d
a na ly s is o f time a d v e r b i a l s and pos si bl e goal expressi ons, n e i t h e r of
which l i e s w i t h i n t he scope o f the present i n v e s t i g a t i o n . Besides,
c onsi de ra bl y more e l a b o r a t i o n would be r eq ui re d if combinations of
var ious time adverbials were taken into consideration, e.g. '/ca0e
mera s tis 6' (ever y day at 6), '4 fores to m in a ' (four times a
month). T he r e f o r e , this brief p re se n t a t i o n is to be understood as
simply i n d i c a t i v e o f the r o l e o f goal and an i n t r o d u c t i o n t o t h a t of
aspect in determing how p r e d i c a t i o n s are understood w i t h r espect to
the ' s t a t e - p r o c e s s - e v e n t ' ( he ncef or th S-P-E) d i s t i n c t i o n .

2.5 Aspect, s t a t e s - p r o c e s s e s - e v e n t s , and Modern Greek motion verbs

Despite the fact that aspect in MG has been the object of probably
more a t t e n t i o n than any ot he r area o f MG l i n g u i s t i c s , t h er e is s t i l l
no uniformity of opinion even on the basic distinctions, i.e. on
whether 'Perfect' should be included t o g et he r w it h 'Perfective' and
' Imperfective' as a t h i r d aspect or not. As most l i n g u i s t s seem to
accept, however, t h a t P e r f e c t is not as c r u c i a l as the ot he r two, the
m a t r i x which i s presented here (a shortened versi on o f t h a t presented
in B a b i n i o t i s and Kontos ( 1 9 6 7 : 1 4 8 ) ) adopts a b i p a r t i t e d i v i s i o n .

Aspect Tense
PAST PRESENT
' Imperfect' ' Present'
Imperfective Durati on 'e tre x a ' 'tre x o '

Non-duration 'Aorist'
'e tre k s a ‘ [ ' t r e k s o Non-Past]
Perfective
(Pf.) Complete Present P e r f e c t a ' ( P r . P f . a ' )
'exo tre k s i'

- 97 -
Future, Subj unct ive and Passive forms are d e l i b e r a t e l y omitted as the
poi nts that have to be made do not require additional i nf or ma ti on
from o t he r tenses, voices or moods (except f o r P e r f e c t b' which w i l l
be mentioned s e p a r a t e l y when it needs to be juxtaposed to Present
Perfect a ' ) . The Pf. non-Past form [ ' t r e k s o ' ] has been added to the
shortened B a b i n i o t i s and Kontos' Table which i s used f o r the A o r i s t
S ubj unct iv e, P e r f e c t i v e Futur e, Pr.Pf. and P l u p e r f e c t .

There is more or l ess general agreement t h a t by using the If. the

speaker sees the verb as r e f e r r i n g e i t h e r to a continuous s i t u a t i o n
in progress, i.e. as a s i t u a t i o n which is ' p r o g r e s s i v e ' or ' d u r a t i v e '
or to a s e ri e s of repeated situations not viewed as a whole, i.e.
'ite ra tiv e ' situations (Mackridge 1985:105, Babiniotis and Kontos
1 96 7 :1 47 ) . A possi bl e combination o f the d i f f e r e n t views expressed
f o r the Pf. would be t h a t by using the ' P f . - n o n - d u r a t i v e ' the speaker
i s viewing the verb as r e f e r r i n g to a s i t u a t i o n which i s momentary or
the d ur a t i o n o f which is not considerd i mpor tant; it i s viewed as a
completed whole, a s i n g l e ' e v e n t ' whose r e p e t i t i o n or d ur a t i o n are o f
no s i g n i f i c a n c e . By using the ' P f . - c o m p l e t e ' ( B a b i n i o t i s and Kontos
ibid.) or the 'Perfect' (Mackridge 1985:116) the s i t u a t i o n i s viewed
as a complete whole, as an event which took place in the past and i s
completed at the time o f u t t e ra n ce (Tzartzanos 1 9 4 6 : 2 7 7 ) . ^

The r elevance of aspectual distinctions in MG for the S-P-E

c a t e g o r i z a t i o n should be q u i t e e v i d e n t . Lyons ( 1 9 6 3 ) , f o r instance,
recognizes the i n t e r co n ne ct i on between aspect and the S-P-E
distinction ( or some form o f t h i s d i s t i n c t i o n ) , on the one hand, and
semantic f e a t u r e s o f i n d i v i d u a l verbs, on the o t h e r . On the basis of
such c onsi de ra ti ons verbs are cat egor i ze d i n t o two main classes: one
comprising 'action-event-state' verbs, the other i n c l ud i n g 'event-
s t a t e ' verbs.

It i s concei vabl e t h a t MGMVs could be c l a s s i f i e d i n t o those which are

generally understood as i n v o lv i n g some feature 'non-durative'/
'momentary', e.g . 'v u ta o j' (dive), 'tin a z o ' (shake up) , ' p /S ao '
(jump), and those which imply ' d u r a t i o n ' , such as: ' x o r e v o ' (dance),
' aneveno' (go up), 'tre x o ' (run). The Imperfective forms of the

- 98 -
verbs belonging to the former class desc ri be iteration of events,
e.g.:

(12) v u t u s e k aBe liy o s ti Balasa (If.+ Past)

s/he dived/was d i v i n g every now and then i n t o the sea.

Their Perfective forms descri be punctual occurrences (example (7)).

The P e r f e c t i v e forms o f the verbs o f the l a t t e r class de scr ibe events
which do, however, have some d u r a t i o n , e.g.:

( P f . + Non-duration + Past)
(13) xo re psa n moni tus s tin p is t a epi b io ores

they danced alone on the f l o o r f o r two hours

It is e v id e nt that (13) is a development rather than a punctual

occurrence owning t o the i nh e re nt semantic feature of dur ati on of
'xo re vo ' (dance).

It has to be noti ced t h a t the verbs of n e i t h e r category can produce

predications with the characteristics of a truly stative situation
( wi th very few pos si bl e exceptions which w i l l be discussed s h o r t l y ) .
The only means a v a i l a b l e w i t h i n the aspectual system o f MG which can
impose a m or e - o r -l e s s s t a t i v e understanding on p r e d i c a t i o n s i n v o lv i n g
such verbs is Perfect b'. Unl i ke Perfect a 7, which refers to
'completed e v e n t s ' , P e r f e c t b' describes the r e s u l t o f an a c t i o n , or
a state "not severed from the time of ut t er anc e" (Babiniotis
1 97 2 : 4 4 ) . This e xp l ai ns why only the l a t t e r tense is incompatibl e
w it h (d) type time a d v e r b i a l s . Consider as an example sentence ( 1 4 ) :

( 14) ine anevasmeni sto bendro

she i s 'gone up' a t / t o the t r e e

she i s on t he t r e e .

Such p r e d i c a t i o n s r e f e r t o the p o s i t i o n r a t h e r than the motion o f an

object, but can be sai d also to i n cor por at e (through the ' r o o t
meaning' o f the verb i nvolved in each case) a r e f e re n ce as to how the

- 99 -
o bj ect a r r i v e d a t the p o s i t i o n i t held a t the time o f u t t e r a n c e . In
my understanding, the resulting p r e d i ca t i o n s receive as s t a t i v e an
interpretation as the i n h e re nt meaning o f the verb can al low for.
Theref ore Pf.b' is the lim iting case o f v er bal s and on the border
wi th a d j e c t i v a l phrases.

Verbs of motion which descri be 'change-of-position', such as

/ ka&ome 2 / ( s i t down), 'ksaplonc>2 ' ( l i e down), 'krem ao' (hang), can be
said t o correspond t o what Lyons (1963) calls 'event-state' verbs.4
Consider as an example ' k s a p l o n o ' :

(15) 7 m arieta ksa plon i

Marietta l i e s / i s lying (down)

can be understood ( i n one i n t e r p r e t a t i o n ) as a punctual occurrence:

the change-of-position is t a k in g place a t the moment of u t t e r a n c e .
In an a l t e r n a t i v e interpretation (15) is durative and homogeneous:
' ksa plon o ' has s t a r t e d a t some uns pe ci fi ed moment before the time of
ut t er ance and may continue a f t e r it. In t h i s understanding (15) is
not u n l i k e (16) which has a number o f ' s t a t e ' c h a r a c t e r i s t i c s .

(16) 7 m arieta ine ksaplom eni (P f.b')

Marietta l i e s / i s lying
M a r i e t t a i s in bed

Not ic e, however, that (16) i s more l i k e l y t o be said o f an i n v a l i d

confined t o bed than ( 1 5 ) . This may wel l imply t h a t (1 5) even in i t s
stative interpretation is l ess stative than (16): Pres.Pf.b'
introduces an element o f permanence as compared t o I f . + d u r a t i o n , all
other t hi ng s being e qual .

We have so f a r discussed the p o s s i b i l i t y o f i d e n t i f y i n g ' e v e n t ' verbs

(i.e . those which can only appear in punctual o c c ur r e n ce s) , ' process-
e ve nt ' verbs (i.e. those which can appear in processes or in
developments), 'event-state' verbs (i.e . those which can appear in
events - punctual occurrences, in e f f e c t - and 'states' o f varying

- 100 -
d egr ee s) . This i deal situation, which matches 'event' verbs with a
semantic feature 'momentary', and 'process-event' verbs wi th a
corresponding feature 'duratio n al', plus an 'in between' cat egor y,
does not take us very f a r . Even w i t h i n the grave l i m i t s arb itrarily
a p pl ie d to this presentation, a great number of cases cannot be
accomwMed in such a frame. F i r s t of a l l , it is not as easy as i t
may seem to differentiate between states, processes and simple
i t e r a t i o n o f punctual occurrences. Consider t he s i t u a t i o n w i t h i n the
area o f d u r a t i o n a l MGMVs. One p a r t i c u l a r verb / t r e m o / ( t r em b le ) does
not have a Pf. form a t a l l . In my understanding i t i s processual par
excellence. It could be also understood, however, as i t e r a t i v e : as
i n v o l v i n g a s e r i e s o f u n i n t e r r u p t e d punctual occurrences. Verbs such
as ' talandevom e' (o scillate), 'eorum e' (sway, swing), which do have
Pf. forms, present a s i m i l a r problem to ' t r e m o ' in p r e d i c a t i o n s where
they appear w i t h If. aspect. An even more i n t e r e s t i n g case can be
exemplified with ' y i r i z o ' (turn). Compare (17) to (18) and ( 1 9 ) .

(17) 7 y7 v i r i z i y iro apo t o n iljo ( I f . + d ur a t i o n + P res. )

the e ar t h tur ns around the sun

(18) o m ilo s y iriz e apo t o n aera apo ( I f .+ d u r a t i o n + Past)

tis 5 os t i s 7

the wi ndmi ll turned/was t u rn i n g

by the wind from 5 t o 7

( P f . + Non-duration + Past)
(19) Y7 r i s e k s a f n i k a k e t o n k i t a k s e
s / h e - t u r n e d suddenly and a t - h i m looked
s/he suddenly turned and looked a t him

There is no need for a detailed a n al y s i s of the specific factors

r e spons ibl e f o r t he d i f f e r e n t understandings o f ( 1 7 ) - ( 1 9 ) here. The
example i s used t o p o i n t out two t h i n g s . Firstly, on the basis o f
the c h a r a c t e r i s t i c s discussed a t l ength in the previous sec ti on (17)
seems more processual than ( 1 8 ) , although i t i s undoubtedly the same
verb and the same aspect in both cases. In my understanding (17) is
very much l i k e a state predication and (18) is not prototypically

- 101 -
processual (like (5), for i ns tance ) but has some of the
characteristics of a development. Once again the borders are not
clear. Secondly, (19) i s a punctual occurrence, as are also possi bl y
most p r e d i c a t i o n s w i t h 'y iriz o ' + Pf, aspect. This i s not the case
wi th either 'x o re v o ' (dance) or 'tre x o 1 (run) which are also
d ur a t i o n a l verbs. On the ot he r hand, the s o l u t i o n o f p o s i t i n g two
different ' y i r i z o ' verbs, a ' d u r a t i o n a l ' one corresponding to (17)
and (18) and a 'momentary' one corresponding to (19) seems r a t h e r
unsatisfa ctory. More importantly, one can get a sim ilar overall
p i c t u r e from a number o f MGMVs, both i n t r a n s i t i v e and t r a n s i t i v e .

2.6 Nature o f the theme

It is i n t e r e s t i n g to n o t i c e t h a t one more f a c t o r c r u c i a l t o the S-P-E

d i s t i n c t i o n can be f a i r l y e a s i l y i d e n t i f i e d in the case o f t r a n s i t i v e
c aus at i ve s; it bears on the c h a r a c t e r i s t i c s o f the object o f these
verbs (and the s ubj ect o f t h e i r corresponding i n t r a n s i t i v e s , i f such
correspondences happen to exist). As a characteristic example,
compare (20) to (21) and ( 2 2 ) ,

(20) ?evaze t o v iv lio sto tra p e zi (If.+ Past)

s/he put/was p u t t i n g the book on the t a b l e

(21) evaze p e tr e J e o s ti beksameni ( I f . + d u r a t i o n + Past)

s/he put/was p u t t i n g o i l i n t o the tank
s/he was pouring o i l i n t o the tank

(22) evaze ta v iv lia sto trapezi ( I f . + d u r a t i o n + Past)

s/he put/was p u t t i n g the books on the t a b l e

(20a) evale to v iv lio sto trapezi ( P f . + Non-duration + Past)

s/he put the book on the t a b l e

- 102 -
(21a) evale p e tre le o s ti beksameni ( P f . + Non-duration + Past)
s/he put o i l i n t o the tank
s/he poured o i l i n t o the tank

(22a) e vale ta v iv lia sto tra p e zi ( P f . + Past)

s/he put the books on the t a b l e

Unl ike (21) where 'evaze' is clearly durational and where the
resulting predication has a number of the characteristics of a
process, (20) can only r e f e r to a s er ie s of 'punctual occurrences'
( or unsuccessful attempts). Moreover, the s i t u a t i o n referred to in
(21) is different from that corresponding to (22), although the
aspectual element o f du ra t io n i s l i t e r a l l y appl ie d to both cases. It
seems t h a t (21) i s a more t y p i c a l case o f a process than ( 2 2 ) , as i t
i s f e l t t o imply more homogeneity. Be t h a t as i t may, (20a) can only
i n volve a punctual occurrence, (21a) is a fa irly typical case of a
development and (22a) may be i n t e r p r e t e d e i t h e r as a development or
as a punctual occurrence (on the assumption t h a t a l l the books were
put on the t a b l e t o g e t h e r , a t one s t r o k e ) .

Similar t hi ngs can be noted for 'sikono skoni ' (I raise dust)
compared to ' sikono to v iv lio apo to patoma' (I lift the book from
the f l o o r ) , or 'v y e n i kapnos' ('rises smoke', i.e. smoke is r i s i n g )
versus ' v y e n i apo to s p iti tis ' (she is coming/going out of her
house). The i m p l i c a t i o n s for the S-P-E distinction of 'mass' vs
' c ou nt ' Ns and 'singular' vs 'p lu ra l' constitute onl y one, fa irly
straightforward, factor. Other factors which are l ess clear and
easily identifiable have a sim ilar effect to that just hi nted at.
For i nstance ' evaze t i roba s to a fto k in ito ' ( s/ he put/was p u t t i n g the
wheel on the c a r) is more likely to behave like (21) than (20),
although the ' f o r m a l ' characteristics of 'ro5a' (wheel) are more
s i m i l a r to those o f ' v i v l i o ' (book) than those o f ' p e t r e l e o ' (o il).

The d e s c r i p t i o n o f the f a c t s w i l l be a r b i t r a r i l y brought to an end at

t h i s p oi n t f o r the reasons a l re ady given in e a r l i e r s ecti ons o f t h i s
chapter: t h i s p re se n t a t i o n is r e s t r i c t e d t o a few poi nts which have
seemed ( a) p a r t i c u l a r l y relevant and (b) fa irly straightforward and

- 103 -
consequently manageable within the space available. These points
have focused on f a c t o r s such as presence or absence and type o f goal ,
aspect and tense and type o f t r a n s i t i v e verb o b j e c t (or in t r a n s i t iv e
verb s u b j ec t ) wi th respect to 'c ount'/'m ass' Ns. The interplay of
such f a c t o r s wi th s p e c i f i c verbs has been shown t o be d e c i s i v e in how
r e l a t e d p r e d i c a t i o n s are understood r egardi ng the S-P-E d i s t i n c t i o n .
I t remains to be seen whether these observations can be used in order
to e s t a b l i s h possi bl e correspondences between a c h a r a c t e r i s a t i o n of
p r e d i ca t i o n s and a c h a r a c t e r i s a t i o n o f s p e c i f i c MGMVs.

2 .7 Concluding remarks on the redefinition of s t a t e s -p ro c es se s -

events

It has been suggested a t various poi nts in t h i s chapt er t h a t n e i t h e r

situations in r e a l i t y nor s p e c i f i c sentences or verbs can be n e a t l y
di s t i n g u i s h e d as possessing a l l and only the c h a r a c t e r i s t i c s u s ual l y
a t t r i b u t e d to s t a t e s , processes and events. They seem, r a t h e r , to be
l ocated along some s o rt o f a S-P-E continuum. On the assumption t h a t
the central/most characteristic poi nts along t h i s continuum can be
identified, s p e c i f i c p r e d i ca t i o n s i n vo l v i n g MGMVs have been o f f e r e d
as c h a r a c t e r i s t i c ' r e a l i z a t i o n s ' o f these poi nt s i n the area o f MGMVs
(examples (4)-(7)). It can be claimed that within this area of
i n v e s t i g a t i o n , p re d i c a t i o n s o f type ( 5 ) : 'tre x i mesa s t o n k i p o ' (s/ he
is r unni ng/ runs within the garden) are instances of prototypical
processes; and p re d i c a t i o n s o f type (7): ' v u tikse sto p o t a m i' ( s/ he
dived i n t o the r i v e r ) are instances of p r o t o t y p i c a l events. It also
seems p l a u s i b l e t h a t processes looked a t 'a p o s t e r io r i ' (i.e. Past-
tensed) w ill be l ess processual, since what they descri be has been
somehow accomplished and its time is more d e f i n i t e than otherwise
(i.e . past). More serious f a c t o r s 'pushi ng' p r e d i c a t i o n s towards an
'event' r a t h e r than a 'pr ocess' understanding are p e r f e c t i v e aspect
and a specification of g oa l, since these i n vol ve an 'anticipated
result' or a 'culm ination'. It is interesting to notice at this
po i nt that verbs which r e q u i r e a goal specification, e.g. 'p iy e n o '

(go) are even less likely to be understood as processual ones than

others which may be equally d ur a t i o n a l and appear in similar
environments. This may e xp l ai n why ( 2 3 ) :

- 104 -
(23) *p iye n e sto s x o lio e p i mi a o ra ( I f . + d u r a t i o n + Past)
s/he was going t o - t h e school f o r one hour

i s f o r some speakers even worse than (24) or ( 2 5 ) :

(24) ? skarfalo ne s tin k o rifi tu vraxu ( I f . + d u r a t i o n + Past)

e p i mi a o r a

• s/he climbed to the top o f the rock

f o r one hour

(25) Imetakomize s to k e n u rjo s p iti ( I f , + durational + Past)

epi b io meres

s/he moved/was moving t o the new house

f o r two days

I t may be t h a t , all ot he r thi ngs being equal, 'pl^eno* incorporates

an 'anticipated resu lt', i.e. is 'inherently' l ess processual (or
more event-like) than ot he r d ur a t i o n a l verbs. It is therefore
suggested t h a t the most p r o t o t y p i c a l i ntances o f a process (in this
area) i n vol ve a motion verb w i t h If. aspect, Pres ent -t ens ed, not
supplemented w it h a goal specification.

The l i n k between the A o r i s t of Greek (both C l a ss i ca l and Modern) and

events seems fa irly wel l established. This is a rather natural
consequence of the fact that non -d u ra ti o n, definite time and an
accomplished situation c h a r a c t e r i z e both the A o r i s t and events. It
has been shown, however, that certain MGMVs al low a collocation of
A o r i s t forms wi th t y p i c a l d u r a t i o n a l a d v e r b i a l s , w h i le others do not,
e.g . ' x o r e v o ' (dance) vs ' s k a r f a l o n o ' ( c l i m b ) :

(26) xorepsan epi bio ores (A orist)

t hey danced f o r two hours

(27) *ska rfa lo sa n e p i bio o re s (Aorist)

they climbed f o r two hours.

- 105 -
It is suggested, therefore, t h a t verbs which behave l i k e 'x o re v o ' in
this respect are more prototypically processual than others which
%
behave l i k e 's k a rfa lo n o '} The equation o f MG A o r i s t wi th absolute
absence o f d u r a t i o n seems unwarranted. It seems, however, plausible
to expect that the most prototypical instances of an event would
i nv ol ve a motion verb i n the A o r i s t wi th a goal and a p o i n t - i n - t i m e
specification, since t y p i c a l events are expected to be t e mpor al ly and
locally restricted. The reason why punctual occurrences are
understood here as more t y p i c a l events than developments is l in ke d
wi th the di scussion of prototypes and the assumption that the most
prototypical i nstances of a category are expected to be maximally
different from those of any ot he r category. Wi thi n the lim its of
this rough S-P-E distinction, punctual occurrences are evidently
f u r t h e r from processes (and s t a t e s ) than developments.

It can be argued t h a t even s i t u a t i o n s c h a r a c t e r i z e d here as punctual

occurrences par e x c e l l e n c e , such as those r epresented by example (7)
'v u tik s e sto pot a m i' ( s / h e dived i n the r i v e r ) , can be said to have
some d u r a t i o n . It is quite conceivable t h a t for some speakers, at
least, 'v u tik s e ' i s v a l i d from the moment the d i v e r ' s f e e t l eave the
ground, to the moment h i s / h e r whole body i s under w a t e r . For othe rs ,
'v u tik s e ' is valid only the minute the diver's body touches the
s urface of the wat er and s t a r t s submerging: t he interval is z er o.
This i s in e f f e c t the i n t e r p r e t a t i o n adopted here f o r ( 7 ) .

It has a lr e ad y been poi nted out that the notion of 's ta te ' is,
s tric tly speaking, incompatibl e wi th motion verbs as such. It has
also been mentioned t h a t P r e s . P f . b ' verb forms are on the borders of
verbal and adjectival phrases. It is as if some pr ope rty is
attributed to the s ubj ec t of t he predication which is regarded as
more or l ess permanent. Comparing If. and P r e s . P f . b ' forms of MG
verbs o f p o s i t i o n (see examples ( 1 5 ) - ( 1 6 ) and subsequent d i s c us s io n) ,
it was suggested t h a t situations described by means o f t he latter
form of the verb are probably viewed as more permanent than
ot her wi se. This i m pl ie s t h a t w i t h i n the area under i n v e s t i g a t i o n the
most p r o t o t y p i c a l instances o f s t at es i nvolve verbs o f p o s i t i o n in a
Pres.Pf.b' form and a t e m p or a l l y unrestricted environment, e.g.
w it hout r e f e re n ce t o the 'in g re ss ive ' stage, ( t he beginning of the

-^(possibly because " x o r e v o ” is not goal orientated).

- 106 -
state), therefore ,ksaplono2 > (lie ) rather than 'aneveno' (go up),
although the l a t t e r verb can also appear in a p r e d i c a t i o n desc ri bi ng
position. I t has t o be emphasized t h a t t h i s i s only t o be understood
within the lim ited area of the verbs under investigation. The
general assumption t h a t I f . forms are more t y p i c a l o f s t a t i v i t y than
Pf. ones is not chal le nged. In fact, 'ks e ro ' (know) which is a
classical example o f a s t a t i v e verb does not have a P f. form a t a l l ;
if r e fe r e n c e has t o be made t o the beginning o f the s t a t e , i.e. the
i ng r e s s i v e stage, a different verb must be used, e.g. 'y n o riz o '
(know), 'k a ta la v e n o ' (unde rs ta nd) , 'maBeno' (learn, find out).

The identification of central instances of sta te s, processes and

events i mpl ies t h a t most p r e d i ca t i o ns are in general t o be understood
not as belonging or not belonging t o one o f these t h r e e c a t e g o r i e s ,
but r a t h e r as being more or l ess p r o t o t y p i c a l ( o r m ar gi na l ) instances
of the relevant categories. This view is chal lenged by an
alternative interpretation of phenomena such as those described
throughout this section. According to this alternative
interpretation a predication may be viewed as c o n ta i ni n g dynamic
elements embedded inside static ones, for i n st an ce , and the S-P-E
s t at us of that predication as a whole is that of the top-most
c o n st r uc t io n (D. Bennett, p .c .). The two views do not seem to me to
be i nc ompat ibl e. An example such as:

( 28) to v iv lio ine valmeno s t o ra fi (P f.b')

the book i s ( p u t ) on t he s h e l f

may be said t o i n v o l v e 'event' m at e r i a l (putting) embedded w i t h i n an

overall stative environment. The suggestion made here i s t h a t (28)
is not a ty pical i nstance of a 'state' precisely because it also
involves 'event' elements (putting). A more s t a t i v e description of
an otherwise s i m i l a r s i t u a t i o n would be:

(2 9) to v i v l i o in e /v ris k e te sto ra fi ( I f . + Pr es .)
the book i s / ' i s found' on the s h e l f

- 107 -
where no r e fe r e n c e is made to the change of state, i.e. the event
which r e s u l t e d in the book's p o s i t i o n . Sim ilarly, an example such
as:

(30) to v iv lio a fto exi b i sto ra fi apo t o 1 91 2 (Pres.P f.a')

t h i s book has ' e n t e r e d ' on-the s h e l f since 1912.
t h i s book has been on the s h e l f since 1912

can be i n t e r p r e t e d as r e pr es en ti ng a s i t u a t i o n even l ess s t a t i v e and

more 'ev en t-like' than that described in (28). Looked at from a
different angle it is again a case of an event embedded i ns id e a
s t a t e (although MG P r e s . P f . a ' i s not u s u a l l y l i n k e d wi th ' s t a t e s ' ) .

2.8 Test frames

This chapter has up to the present poi nt presented the view that
whole p r e d i c a t i o n s can be c h ar a c t e r i z e d as more or l e s s p r o t o t y p i c a l
instances o f s t a t i v e , processual and ' e v e n t ' s i t u a t i o n s depending on
a number of d i f f e r e n t elements present in them. Some o f the r e l e v a n t
elements have been i d e n t i f i e d (aspect, g o a l, etc.). It is thought
t h a t c o m p a t i b i l i t y o f d i f f e r e n t MGMVs w it h d i f f e r e n t combinations of
some o f these elements may gi ve a p i c t u r e (however i na cc u ra t e) o f the
relative 's ta tiv ity ', 'processuality' and ' e v e n t - l i k e n a t u r e ' o f the
verbs under investigation. Clearly the number of 'com patibility
possibilities' is great. For reasons o f space and 'manageability' of
the m a t e r i a l , however, only f i v e such p o s s i b i l i t i e s are considered
and used jo in tly as a heuristic. Far from being a full-blooded,
categorization, the resulting schema which appears in List III
consi sts simply o f ' f r a m e s ' used as tes t- env ir onme nt s f o r some MGMVs.
P o si t i o n verbs are not t e s t e d , n e i t h e r are ' s t a t i v e ' verb-forms ( e . g .
Perfect tenses). The reason is rather obvious: if a formal
d i s t i n c t i o n i s to be made anywhere, "it is more l o g i c a l f o r i t to be
made within descriptions of dynamic situations than within
descriptions of s t a t i c situations" (Comrie 1 9 7 6 : 51 ) .

The ' fr ame s' used are as f ol l o w s:

- 108 -
I. If. aspect + Present + [ e p i m i a . . . o r a ] P P
f o r one . . .hour
(Continuous motion)

II. Aorist ( i . e . P f . + ' n o n - d u r a t i o n ' + Past) + [ e p i m i a . . . o r a ] P P

(Motion understood as continuous but completed ( d e f i n i t e t i m e ) )

I I I . A o r i s t + [ a p o NPlqc s e Np|_0C mesa se w / a . . . o r a ] P P

from NPloc to NP|_q c in o n e . . . h o u r
(Motion understood as r e s t r i c t e d both l o c a l l y and te mp or a ll y )

Each verb i s put in one or more o f the r e s u l t i n g columns B, C, D, E, F,G

according to which frames it is compatible w i t h . Thus, column B
contains those verbs t e s t e d which are compatible with all thr ee
frames (I, II and III). Column C contains the verbs which all ow I
and I I I but block I I (hence the box corresponding t o II is black).
Column D verbs a l l ow only frame III. The remaining two columns (E
and F) contai n verbs which normally block a l l three possibilities.
Their difference lies in t h a t w hi le E verbs may appear in frames I
and II if the implication is 'i t e r a t i o n ' rather than 'continuous
motion' (which the frames are meant to i m pl y ) , F verbs do not allow
II at all ( i . e . even if we wish to attempt an 'ite ra tiv e '
u n d e rs t an d i ng ) .

It is assumed t h a t the f u r t h e r t o the l e f t in L i s t I I I a verb appears

the more processual it is, and the f u r t h e r it appears to the r i g h t ,
the more ' e v e n t - l i k e ' it is. It can be said t h a t roughly speaking E
and F verbs are 'event' verbs. P r o t o t y p i c a l l y processual MGMVs are
also excluded from the test and constitute column A. These are:
' trem o' (tremble), 'p e rife ro ' (t ake s.th. here and there),
'p e rife ro m e ' and 'triy iriz o ' both meaning 'roam around',
'k ik lo fo ro j (circulate). The reason i s that they e i t h e r do not
have a Pf. form a t a l l ('tre m o ') or, i f they do, it is of r e s tric te d
use, so they cannot be r e a l l y submitted to t e s t s making use of the
Aorist. Besides, as they are i ncompatibl e with the noti on of a
'journey' (i.e . 'fr om one point to another'), t hey cannot be
submitted t o the t e s t f o r frame I I I . I t seems i n t u i t i v e l y c o r r e c t to

- 109 -
posit 'tre m o ' as the most p r o t o t y p i c a l i nstance of a processual MGMV
and consider the others i n between c at egor i es A and B. An a d d i t i o n a l
candidate f o r such an in between category which also blocks III (for
reasons ot he r than the combination of Aorist plus a 'journey'
s p ec i f y i ng PP) is the set of verbs: ' a n a k a te v o jp -o m e j^jp ' (stir,
to ss), 'a n a ta ra s o jft-o m e jN jft' ( s t i r up, shake), 'ta ra k u n a o jp -je m e j^ jp '

(shake up). The verbs o f t h i s set present a speci al problem, namely

th a t it is not immediately obvious whether i t is 'continuity' or
'iteratio n ' t h a t they i mply. A s i m i l a r problem i s present in verbs
appearing a t the ot he r end o f t h i s ' s c a l e ' o f frames ( c a t e g o r i e s ) and
i s discussed in t h a t connection under Comment 12. Notice t h a t each
column i ncludes also the numbers r e fe rr in g to the comments which
accompany s p e c i f i c sets o f verbs or i n d i v i d u a l verbs. These comments
on L i s t I I I f o l l o w immediately.

2 . 9 Comments on L i s t I I I

1, 'p e rik ik lo n o jp -o m e j^ jp ' (encircle) can pass II, but wi th the

implication: 'performed the act and then stayed there for
one...hour'. Since t h i s i s c l e a r l y not the intended i m p l i c a t i o n ,
it appears in category C. The same a p pl i e s to a number o f ot he r
Vs belonging t o d i f f e r e n t c at egor i es (e.g. 'k a d iz o jp -k a B o m e jp jp '

(sit)) and w i l l not be marked every time i t occurs as i t does not

affect the issue at hand,

2a. The main p oi n t in r e l a t i o n to t h i s set o f Vs is t h a t i f the goal

is s p e c i f i e d they move to category D, i . e . become l ess processual
(and consequently more 'e v en t-lik e '). Differences between
individual Vs as to the type of goal expression each of them
allows are o f no concern here ( e . g . 'tre x o sto p e rip te ro ' ( I run
to the k i osque ), vs * ' p e r p a t a o sto p e rip te ro ' (I walk to the
k i osque). Obviousl y, the same environment would al low the A o r i s t
provided we have NP|_qc instead o f g oa l, i.e. a sp ec ific atio n of
the place i n / a t / w i t h i n which the motion described by the V takes
pl ace: e.g. 'tre x i ston k i p o e p i mia ora' is fi n e only i f it is
understood as ' s / h e r u n s / i s running i n / w i t h i n the garden f o r one
hour'.

- 110 -
2b. This set i s again c h a r a c t e r i z e d by a t r a n s p o s i t i o n t o the r i g h t
(from C to D ), thus giving rise to rather 'event-like'
p r e d i c a t i o n s , w i t h a goal s p e c i f i c a t i o n , e . g . :

* anevenume s t i n k o rifi t u k a b a n a rju e p i mi a o ra

we go/ ar e going to the top of the b e l l - t o w e r f o r one hour.

3. One o f the major problems of this categorization, and a c l e a r

example of i ndetermi nacy, in my view, is exemplified by
' ska rfa lo n o ' ( c l imb ) and 'p ib a o ' (jump). Both of them could
concei vabl y (in ot he r words, for some speakers) behave like B
cat egor y/ fr ame Vs i f they are understood as implying 'continuous
physical activity/exercise' in a c onte xt concatenati ng d i f f e r e n t
' manner -of-mot ion' instances o f physical exercise, e.g . 'etrekse
epi m ia ora, ska rfa lose epi m is i or a, p e rp a tis e epi b io ores,

p ib ik s e epi m is i ora, ...' (he ran f o r an hour, climbed f o r h a l f

an hour, walked f o r two hours, jumped f o r h a l f an h o u r , . . . ) .

4. I t seems i n t e r e s t i n g t o compare ' v u t a o ' ( d i v e ) t o 'p ib a o ' (jump);

it appears t h a t the only reason why the former could not appear
in category E i s t h a t , although i t can be also i n t e r p r e t e d as an
i nstance of ' phy s ic a l exercise', it i s more momentary than the
former and even l ess homogeneous,

5. This set allows C environments (C frame) wi th an NPlqc ( as above,

s p ec i f y i n g the place where/on/within which the motion takes
p l ace) or w i t hout any l o c a t i v e s p e c i f i c a t i o n ( e . g . ' a nevenume e p i
mia ora' (we g o / a re going up f o r an h ou r ) , 'anevenume ti skala
e p i mia o r a ' (we g o / a re going up the s t a i r c a s e f o r an hour).

6. As a l re ady discussed in connection w i t h the n om i n al i z at i on s t e s t ,

these Vs present a speci al problem. To s t a r t with, 'm e tafe ro '

( c a r r y / t r a n s p o r t ) has no d i s t i n c t p e r f e c t i v e aspect (hence A o r i s t

and I mper fec t are m or p ho l og i ca l ly i d e n t i c a l ) . T he ref or e only the
passive form ( ' m e t a f e r o m e ' ) can be submitted t o the t e s t s . They
also seem to block II if goal is specified, but I have been
unable to get. agreement from n a t i v e speakers as t o whether they

- Ill -
 also block I or not. It seems to me t h a t the interpretation
corresponding t o English ' c a r r y ' behaves more ' p r o c e s s u a l l y ' than
the 'transport' use. In either case they both seem to become
more 'e ven t-lik e' wi th an e x p l i c i t specification o f goal (e.g.
* ' 7 v a lits a m eta fe rQ ike s to staQmo e p i m i a o r a ' ( t he s ui tca se was
c a r r i e d to the s t a t i o n f o r one h o u r ) ) .

7. A r a t h e r i n t e r e s t i n g case is e xe mpl i fi ed wi th 's ik o n o jp -o m e jp jp '

(ra is e /lift, rise). The commonest understanding o f these verbs

seems to be the one focusing on the source, e.g. 'siko se to

v iv lio apo to patom a' ( s /he lifted the book from the floor).
This is the most restricted interpretation t e m po r a ll y (and
perhaps also l o c a l l y ) o f these verbs; hence they are l i s t e d under
F, t o g e t h e r w it h t y p i c a l l y 'event' verbs, wi th the proviso t h a t
III is inapplicable, s tric tly speaking, since it exp licitly
r e f e r s to both source and g o a l. If, however, the focus moves to
the goal or the 'journey' o f the theme, these verbs behave much
less like 'event' verbs and are l i s t e d under C and D, e.g. 'o
yeranos s i k o n i to pjano s to te ta rto patom a' (t he crane is r a i s i n g
the piano t o the f o ur t h f l o o r ) .

8. A vast number of the verbs appearing in this lis t change

category depending on the nature o f the theme(moving o b j e c t ) , a
general issue al re ady discussed b r i e f l y in the r e l e v a n t section
which w i l l be expl ai ned t h e o r e t i c a l l y at g r e a t e r l ength in the
next one. What needs to be noted here i s t h a t if the theme is
'mass' or ' p l u r a l ' , the verbs marked f o r Comment 8 in category C
move to category B, w hi le those marked f o r the same comment in
c at e gor i es E and D move to C. An a d d i t i o n a l but r e l a t e d po i nt
has be made in connection wi th 's te ln o ' ( send). Not ice the
d i f f e r e n c e between (a) and (b) below:

(a ) *s te ln o y rama s t i y e r m . n ia e p i mia o ra

I send/am sending a l e t t e r to Germany f o r one hour

(b) s t e l n o m i n i m a me t o n a s i m a t o e p i mia o ra

I am sending a message by r adi o f o r one hour.

- 112 -
 On the basis o f the (a) type of examples ' s t e l n o ' i s l i s t e d under
F, w h i l e the (b) type o f sentences w i t h the same verb poi nt to
t he possibility of its moving to category C. It is mentioned
s e p a r a t e l y in order to i n d i c a t e t h a t ' c o un t ' vs 'mass' has to be
further analysed in connection wi th certain entities (e.g.
'm in im a ' is gr ammati call y 'count'). A sim ilar point has been
raised in connection with ' v a zo ti roba' (I put/am p u t t i n g the
wheel (on the c a r ) ) in section 2 . 6 . The whole issue is taken up
i n the f o l l o w i n g s ec ti o n .

9. The case o f verbs such as ' y i r i z o j (turn) has al re ady been

discussed in the secti on on aspect and i t should be q u i t e c l e a r
by now t h a t such verbs can cover almost the whole range o f S-P-E
possibilities. They appear here under F w it h the proviso t h a t
this a ppl ies only to their 'non-durational' interpretation. A
sim ilar situation is presented by 'k u n a o j' (move) which (quite
ch aracteristically) can appear in almost any category depending
on a number o f d i f f e r e n t f a c t o r s . A more serious problem ar is es
with 'kunao?' (move) which belongs tog e th e r wi th a number o f Vs
implying motion w it ho ut c h a n g e - o f - l o c a t i o n - i n p a r t i c u l a r (a)
'ta la n d e v o jp -o m e jN Y p ' (o s c illa te ), 'eorum e' (sway), (b) 's k iv o '
(bend), 'yerno' (lean) and (c) 's a le v o j^ jp ' (stir, move
s lig h tly). All these w i l l normally block a ' j o u r n e y ' expression
(as they are not c h a n g e - o f - l o c a t i o n Vs); those in subset (a) pass
I 'o ffic ia lly ' but i t i s almost impossible t o decide whether we
are d e al i ng wi th ' conti nuous' and 'homogeneous' motion, or with
iteration . Neither 's a le v o jp jp ' (s tir), nor 'k u n a o jp ' (move)
a ll o w f o r an i t e r a t i v e understanding i n II environments, u n l i ke
the verbs of subset (a). These last verbs resemble the ones
a l re ady mentioned a t the beginning of these comments: 'a n a k a te v o '

( s t i r , toss), ' a n a t a r a s o ' ( s t i r up, shake) and ' t a r a k u n a o ' (shake

up) which pass both I and II, but f o r which it is once again
d iffic u lt to decide whether 'continuity' or 'iteratio n' is
i mp l i e d . Hence the main reason f o r c l a s s i f y i n g only ' ta lan de vo '
(oscillate) and ' eorume' (sway) under E is that they clearly
i n vol ve a regular type of motion, so that the 'ite ra tio n '
i n t e r p r e t a t i o n seems more n a t u r a l .

- 113 -
10. Not ic e, that ' p e t a o^ ' (throw) w i l l accept a ' j o u r n e y ' expression
(from NPlqc t o NPlqc ) but the time expression f o l l o w i n g i t counts
the time which f ol l ow s the act o f throwing.

11. These Vs [ ' x o r e v o ' , 'x o ro p ib a o ' , 'k ilje m e ') are i ncompat ibl e wi th
e i t h e r a goal or a ' j o u r n e y ' expressi on, but t h i s does not a f f e c t
the issue a t hand and w i l l be t h e r e f o r e di sregarded here.

12. It has seemed better to consider 'g re m iz o ' (pull down)

separately. ' Na tur e of theme' is again at i ssue, though in a
sense different from the 'mass' vs 'count' noun distinction.
Not ice the d i f f e r e n c e between (a) and (b) below. Example ( a ) :

(a) gremizo to s p iti epi b io meres

I am p u l l i n g down the house f o r two days

I have been p u l l i n g the house down f o r two days

i nvolves c l e a r l y a C category verb, whi le ( b ) :

(b) * g r e m i z o t o n a n d r o p o apo t i n k o rifi tu lo fu epi b io lepta

I ' t h r o w down' the man from the top o f the h i l l f o r two
minutes

shows t h a t we are in the area o f category F ver bs. Naturally,

(b) would be acceptabl e i f t he d u r a t i v e expression were absent.
The d i f f e r e n c e between the ( a ) and (b) occurrences o f 'g re m iz o '

seems to lie in the nature of the theme; t he different

understandings of the two motions i nvolved can be viewed as a
conti ngent fact, i.e. as depending on the type of o bj e ct
undergoing 'g re m iz o '. P o si t i n g two d i f f e r e n t verbs seems r a t h e r
unwarranted. The verb appears in category C r a t h e r than F simply
because t he use e x e m p l i f i e d in (a) i s much commoner than the one
e x e m p l i f i e d in ( b ) .

13. The c r u c i a l f a c t o r f o r the extreme 'e ven t-like ' na tur e o f these
Vs seems t o be ' s p e c i f i c a t i o n o f sour ce '.

- 114 -
14. Not ice t h a t ' o b i y o j ' (drive, e.g. a c ar ) and ' o b / y o ^ ' (l ead s . o .
s .wh. ) are understood here as d i f f e r e n t verbs although they are
obvi ousl y related, otherwi se 'o b iy o j' would not have been
included.

15. The case o f 'y lis t r a o ' is different. For the moment, the two
main i n t e r p r e t a t i o n s o f the verb are marked a s ' y l i s t r a o j ' ( s l i p )
and ' y l i s t r a c > 2 ' (slide). I t w ill be shown (on the basis o f the
t e s t r e s u l t s discussed in Chapter 5) t h a t t h e r e i s good reason to
believe that ' s l i p ' i s by f a r the most immediate i n t e r p r e t a t i o n .

- 115 -
 Notes on Chapter 2

1. The only apparent exceptions t o t h i s are wishes expressed in the

form o f an I m p e r a t i v e , e . g . Sleep wel l or Have a ni ce t i m e .

2. The only method I am not prepared to discuss is d e f i n i n g HAPPEN

on the basis of the ' state-process-event' distinction and then
using it for further testing ot he r verbs, as I find it rather
circular.

3. None of these accounts of MG aspect seems to take notice of

Comrie's d i s t i n c t i o n between 'compl ete' and ' c o m p l e t e d ' , but as
i t seems an i mportant one i t w i l l be presented here in f u l l :

"A very f re que nt c h a r a c t e r i s a t i o n o f p e r f e c t i v i t y is t h a t i t

i n d i c a t e s a completed a c t i o n . One should note t h a t the word
a t issue in t h i s d e f i n i t i o n i s ' c o m p l e t e d ' , not ' c o m p l e t e ' :
d e sp i te the formal s i m i l a r i t y between the two words, there
i s an i mportant semantic d i s t i n c t i o n which tur ns out to be
c r u c i a l in di scussing aspect. The p e r f e c t i v e does indeed
denote a complete s i t u a t i o n , w i t h beginning, middle, and
end. The use o f ' c om p l et e d' , however, puts too much
emphasis on the t e r m i n a t i o n o f the s i t u a t i o n , whereas the
use o f the p e r f e c t i v e puts no more emphasis, n e c e s s a r i l y , on
the end o f a s i t u a t i o n than on any o the r p a r t o f the
s i t u a t i o n , r a t h e r a l l pa rt s o f the s i t u a t i o n are presented
as a s i n g l e whol e. " (Comrie 1976:18)

4, A reasonable interpretation o f the st at us o f such verbs is that

adopted by Comrie for 's it' namely t h a t it can in general be
either stative ("be s i t t i n g " ) or 'ingressive' ("adopt a sitting
position") (1976:20).

- 116 -
 3. THE 'CAUSATIVITY - AGENTIVITY' CATEGORIZATION AND
ITS APPLICATION TO MOTION VERBS

3.1 How separate can the notions of ' c a u s a t i v i t y ' and ' a g e n t i v i t y '
be kept?

Despite years of phi l osop hi ca l and l i n g u i s t i c work on c a u s a t i v i t y

major problems r e l a t i n g to t h i s issue remain unsolved. This anal ys is
d o e s. n ot pur por t to t a c k l e them. But 'cause' being undoubtedly one
of the main c l a s s i f y i n g semantic f e a t u r e s for verbs, i t has to be
examined in d e t a i l in an a na ly s is o f a verbal domain such as t h a t of
MGMVs.

In order to investigate the behaviour o f MGMVs i n r e l a t i o n to the

properties in quest ion, i.e. c a u s a t i v i t y and a g e n t i v i t y , a number o f
issues have to be discussed:

the e xt e n t to which the notions of ' c a u s a t i v i t y ' and ' a g e n t i v i t y '

can be kept separ ate;

the relation of causativity to tra n s itiv ity and the notion of

'p a tie n t';

t he r e l a t i o n o f a g e n t i v i t y to the d i s t i n c t i o n between d i r e c t
and i n d i r e c t c a us a t i ve s .

Two c a t e g o r i z a t i o n s o f MGMVs w i l l be attempted, one on the basis o f

c a u s a t i v i t y a t the end o f the f i r s t secti on and a separ ate one on the
basis of agentivity at the end of the final section. These
categorizations require a prior c la rific a tio n of the notions of
' cause' and 'e ffect' and an understanding of 'ag en tivity' as a
c lu s te r of properties ( r a t h e r than as a s i n g l e f e a t u r e ) . I t w ill be
shown that just as for the S-P-E distinction (discussed in the

- 117 -
previous chapter), whole p r e d i c a t i o n s w ill have to be examined for
presence or absence o f the relative properties and t h a t individual
verbs can be c a l l e d 'a gentive' to a g r e a t e r or l e s s e r e x t e n t (i.e.
t h a t presence or absence o f t h i s pr ope rt y i s not a yes-no q u e s t i o n ) .
Although c a u s a t i v i t y seems to be more obvi ousl y a yes-no m a t t e r , the
possibility that, i t , t o o, might have a g r a d i e n t na tur e seems worth
investigating.

Probably one o f the l e a s t c o n t r o v e r s i a l and most concise accounts of

a caus at ive c o n st r u c t i o n i s t h a t o f f e r e d i n Shibatani (1976a:239-40),
C a u s a t i v i t y i s s t a t e d as a r e l a t i o n between two events which holds i f
the f o l l o w i n g c o n di ti o ns are s a t i s f i e d :

"a. The r e l a t i o n between the two events i s such t h a t the speaker

b e li ev es t h a t the occurrence o f one event, the 'caused
e v e n t ' , has been r e a l i z e d at which i s a f t e r t j t the time
o f the 'causi ng e v e n t ' .

b. The r e l a t i o n between the causing and the caused event is

such t h a t the speaker b e li e v e s t h a t the occurrence of the
caused event is whol ly dependent on the occurrence of the
causing event; the dependency o f the two events here must be
t o the e x t e n t t h a t i t allows the speaker t o e n t e r t a i n a
c ou nt e r f a c t u a l i n f e r en c e t h a t the caused event woul dn' t have
taken place a t t h a t p a r t i c u l a r time i f the causing event
hadn' t taken p l a ce , provided t h a t a l l e l s e had remained the
same."

This is e n t i r e l y i n keeping w it h the common phi lo so p hi ca l view t h a t

in order to a s s er t that an o b j e c t ' s a ct ions caused some event to
happen, a speaker must b e l i e v e that the event happened and t h a t it
would not have happened on t h a t p a r t i c u l a r occasion i f t he o bj e c t had
not acted and a l l e l s e had been the same ( c f , also S t a l n a k e r 1968,
Dowty 1972a, M i l l e r and Johnson-Laird 1976).

This is perhaps as far as agreement on this issue goes. The

relationship between causativity and agentivity is philosophically
controversial and l i n g u i s t i c discussions o f caus ati ve verbs ignore
the distinction between causers and agents in practice, even when
they recognize its e xi s te n ce in t h e or y . Before prov idi ng specific

- 118 -
examples o f t h i s c la im i t i s worth d i gr e ss i ng b r i e f l y t o look f o r an
ex pl a na t io n of the confusion.

It i s o ft e n assumed t h a t the perceptual mechanism f o r causation is

i nnate and t h a t v is u al causation provides the p oi n t of origin for
later conceptual r ef ine me nt s . M iller and Johnson-Laird (1976:489
ff.) consider also the possibility that the perceptual predicate
CAUSE may originate w it h the intentional movements of the child,
g e n e r a l i z e to ot he r persons and f i n a l l y to inanimate o b j e c t s . Lyons
(1977:482 ff., drawing on P ia ge t ) accepts a sim ilar account o f the
child's conception o f causality originating in action. If this is
the case, the concept o f causation may be the r e s u l t o f f a i r l y l ong-
lasting cognitive development along the lines already mentioned:
'ego' i s r eplaced by ' p e r s o n ' ; o t he r peopl e' s i n t e n t i o n s are i n f e r r e d
through t h e i r per cei ved behaviour; and the end o f t h i s process is the
average adult's noti on of intention and the related notions of
animacy and agency. If cause and i n t e n t i o n are so c l o s e l y related
ontogenetically we have an e xp l ana ti on for the confusion between
c a u s a t i v i t y and a g e n t i v i t y , since t he re has never been any doubt t h a t
intentionality is at least one of the main p r o p e r t i e s of an agent
(and on most accounts i t i s understood as an agent ' s sol e p r o p e r t y ) .

Be t h a t as i t may, it is also well known t h a t although it takes at

least ten years for an average child to get his/her ideas about
causation and animacy sorted out , s/he i s a competent speaker o f h i s /
her language long befor e t h i s is accomplished (M ille r and Oohnson-
La ir d 1 97 6 :4 91 ) . The f a c t also t h a t agents are i nter cha nge abl e wi th
o t he r kinds o f causers ( e . g . The umpire stopped pl ay or Rain stopped
pi a y) is brought as evidence for the exi st ence of a "na tur al and
perhaps u ni ver sa l tendency to i d e n t i f y c a u s a l i t y w it h agency" (Lyons
1 9 7 7: 4 90 ) .

In s h o rt , c o l l a p s i n g the notions o f a g e n t i v i t y and c a u s a t i v i t y i s , to

some extent, ju stifiab le. It seems, however, that a clear
understanding of the poi nts at which they intersect and those at
which they are distinct is necessary befor e any attempt at
c a t e g o r i z i n g verbs along these dimensions i s made. In p a r t i c u l a r i t

- 119 -
is important to recogni ze that causativity hinges on the result
(effect) or the s i t u a t i o n o f the r e c i p i e n t o f an a c t i o n , not on the
qualities of the causing event as such. The extent to which the
causing event i nvolves an agent, i.e. an animate, vo litional,
r espons ibl e entity who is in control of the situation and has
authority over it w ill be decisive in c ons ide r at ions of the
'a g entivity' of a predication (and f i n a l l y o f the verb i t involves).
Although these observati ons are not contradicted in t he or y in any
ana ly s is that I know of, it can be shown that specific
categorizations of predications (and verbs) in relation to these
basic p r o p e r t i e s c o n f l a t e them i n p r a c t i c e and are consequently more
confusing than illum inating. The cle arest examples of this
conflation are probably Lyons (1977) and M iller and Johnson-Laird
(1976) as both works provi de excellent theoretical accounts of the
relevant notions and r e l a t i o n s . They w i l l be t h e r e f o r e considered
here in some d e t a i l .

Lyons' ( 1 9 77 : 4 9 1 - 4 ) categorization of verbs in relation to

causativity results in a trip a rtite division into 'fa c titiv e ',
'operative' and 'o p era tive -fac titiv e' verbs. In his system
'fa c titiv e' verbs are sai d to denote a situation where a cause
produces an e f f e c t :

(a) PRODUCE (CAUSE, EFFECT)

'Operative' verbs are sai d to denote an o pe rat io n performed by an

agent and a f f e c t i n g a p a t i e n t :

(b) AFFECT (AGENT, PATIENT)'

' O p e r a t i v e - f a c t i t i v e ' verbs are c h a r a c t e r i z e d by a combination o f the

elements appearing in (a) and ( b ) :

(c) PRODUCE (AGENT, EFFECT)

- 120 -
All t h r e e schemata are said to be r e l e v a n t in "paradigm i nstances of
a gent ive situations (i.e . those in which the a c t io n results in a
change in the physical c o nd i ti on or location of the patient)"
( i b i d . :491). A distinction i s drawn, however, between i nstances of
'k ill' where the cause i s a first-order nominal ( John k i l l e d B ill)
and those where it is a second-order nominal ( Excessive d r i n ki n g
k ille d B i l l ). In the former case ' k i l l ' i s understood as c l o s e r to
(c) while in the l a t t e r it is a 'fa ctitive' verb, i.e. it relates
directly to (a). Finally, since the p ro p o s i t i o n expressed by fx

killed y ' can also be understood as saying t h a t x di d something to y

"it can be understood as an i nstance of AFFECT (AGENT, PATIENT)"
( i b i d . :492), i.e. 'k ill' can be also an ' o p e r a t i v e ' verb. This sort
of categorization results in two d i s t i n c t understandings of 'k ill':
an agent ive one corresponding to (b) and (c) and a non-agentive
i n t e r p r e t a t i o n corresponding to ( a ) . Besides, in so f a r as a f f e c t i n g
a patient i s not e q u i va l e n t to producing an e f f e c t , an agent ive but
non-factitive understanding is also p r ed i ct ed as a possibility for
'k ill'. It seems to me t h a t the f a c ts in connection w it h 'k ill',
causativity and a g e n t i v i t y can be accounted for in a simpl er way,
provided the two notions are kept separate. What we know to be tr ue
of a ll instances o f 'k ill' is t h a t it is a caus at ive ver b. He was
killed in a bomb explosion and He was k illed during the last
earthquake i nvolve perfectly lite ra l instances of 'k ill'. An
e x t er n al cause is unequi vocal ly requi red (unlike in the case of
'd ie ', for instance). In this r e spe ct , i.e. in anything that
concerns c a u s a t i v i t y , both examples belong t o g e t h e r w it h John k i l l e d
Bill where unlike either a genuine agent is i n v o lv e d. The ref or e,
'k ill' i s undoubtedly a c a us a ti v e, but i t i s not e q u a l l y c e r t a i n t h a t
it is also 'completely' a ge nt i ve , since it does not necessarily
require an animate volitional causer, being r e spons ibl e for the
situation, having a u t h o r i t y and control over i t . Such f a c t o r s (which
w ill be discussed separately in the relevant section) have t o be
taken i n t o c ons id e ra t io n only in order t o decide the e x t e n t to which
'k ill' is an agent ive verb and i t i s easy t o see t h a t it w ill not
fe ature at the top of an a g e n t i v i t y scale, unlike a verb such as
' m u r d e r ' , f o r i ns tan ce .

- 121 -
It looks as i f a trip artite categorization i s a t the same time both
redundant and i n s u f f i c i e n t to account f o r the f a c t s o f ' k i l l ' . For,
besides the examples a lr e a d y discussed, the same verb may r e ce i ve
additional interpretations in cases like: John killed B ill by
acci dent and John k i l l e d B i l l in order to s t e al his c o c a i n e . None o f
these i n t e r p r e t a t i o n s hinges on the e f f e c t , the r e s u l t o f the causing
event, i.e. none o f them a f f e c t s the 'causativity' i ss ue . They a l l
bear on the e x t e n t t o which the causer is an agent, i.e. they a l l
depend on p r o p e r t i e s o f the causing event and a f f e c t t he ' a g e n t i v i t y '
i ssue.

Consider also the case o f 'h it' and Lyons' explanation o f why the
verb i s not ' o p e r a t i v e - f a c t i t i v e ' . His s p e c i f i c example i s John h i t
B ill and t he e xp l an a ti o n i s worth l ooking i n t o r a t h e r c a r e f u l l y :

"we can, o f course, say t h a t in so f a r as some change i s wrought

in the c o ndi ti on o f B i l l , John's a ct io n r e s u l t s i n a new s t a t e .
But English does not provide us w it h a monovalent p r e d i c a t o r
denoting such s t a t e s " ( i b i d . : 4 9 2 ) .

This means that w h i le in the case of 'k ill' there exists an

i n t r a n s i t i v e verb ' d i e ' which r e f e r s t o the r e s u l t brought about, no
such i n t r a n s i t i v e verb e x i s t s i n the case o f 'h it'. T he r e f o r e , for
Lyons 'e ffe c t' equals ' l e x i c a l i z e d new state' and consequently
'absence o f lexicalized new s t a t e ' i mpl ies 'no e ffe c t'. In shor t,
'h it' i s not ' o p e r a t i v e - f a c t i t i v e ' f o r the same reason t h a t i t is not
just 'fa c titiv e' either, presumably ( si nce 'fa c titiv es ' also r e q u i r e
that an e f f e c t be produced). What is 'h it' then? Lyons does not
te ll us how i t i s t o be c h a r a c t e r i z e d , but we may attempt t o c a l l it
an ' o p e r a t i v e ' ver b. 'Operatives', however, i nv ol ve a ' p a t i e n t ' . In
an example like John hit the wal l w it h his fis t and hur t h i mse lf
nothing happens t o the w a l l . It i s not t h e r e f o r e e v i d e n t t h a t the
prototypical
wal l is indeed a / ^ ' p a t i e n t ' . Is it consequently t r u e t h a t we do not
know anything about 'h it' in connection w it h agentivity and
causativity? The answer i s n e g a t i v e , because we do know t h a t in both
examples ' h i t ' i s an agent ive ver b. The e x t e n t to which the subject
possesses t he r e l e v a n t c h a r a c t e r i s t i c s t o q u a l i f y f o r an ' a g e n t ' can
be worked out ( i n a s p e c i f i c way which w i l l be presented l a t e r ) , but
the fact remains that some of these relevant characteristics are

- 122 -
always present in all instances of 'h it'. It can be argued, of
course, t h a t ' p a t i e n t ' is t o be de fi ned as the e n t i t y which i s on the
r e c e i v i n g end o f an a ct ion and in t h a t case the wal l in the ' h i t the
wall' example is a patient. Not ice, however, that the c r u c i a l
element once again is 'ac tio n '; i f t h a t notion i s also absent we are
saying nothing more about ' h i t ' than t h a t i t i s a t r a n s i t i v e verb and
that is not much of a categorization. Clearly 'action' i mplies
presence o f an ' a g e n t ' . Hence, i r r e s p e c t i v e o f the exact d e f i n i t i o n
of 'p a tie n t', the decisive factor, which remains constant in the
examples of 'h it' just discussed, is agentivity, A categorization
which mixes up elements c r u c i a l f o r c a u s a t i v i t y and elements c r u c i a l
f o r a g e n t i v i t y seems to miss the p o i n t .

In the case o f MVs presence or absence o f c a u s a t i v i t y seems t o be a

f a i r l y s t r a i g h t f o r w a r d m a t t e r , provided one s t i c k s to the t r a d i t i o n a l
notion o f a c a u s a t i v e , r e q u i r i n g causative verbs to be t r a n s i t i v e s ,
for a start. Motion verbs seen in t h i s l i g h t would be causati ve i f
they descri be the motion o f an o b j e c t which can be s a f e l y a t t r i b u t e d
to the interference (or action) o f some o t he r e n t i t y . This simple
view is not as w i d el y adopted in p r a c t i c e as one might have expected.
M iller and Johns on- La ir d' s (1976) characte rization of MVs with
r espect to c a u s a t i v i t y r e s t s on an a l t e r n a t i v e view which seems to
also add to the confusion o f the notions of causativity and
agentivity. On their account, if the moving o bj e ct and the one
r esponsi bl e f o r i t s motion are one and the same, the verbs de sc ri bi ng
the motion in question are also understood as ' c a u s a t i v e s ' . Clearly
these verbs are not transitives any longer but intransitives of
motion; moreover the issue of coreferentiality (i.e . whether the
moving o bj e c t and the one r espons ibl e f o r i t s motion are one or not)
can only be r a i s e d in those cases where t he moving o b j e ct is an
agent. In the understanding of the proponents of this view, the
moving o bj e ct must in such cases be capable o f v o l u n t a r y a c t i o n . As
an i l l u s t r a t i o n o f t h i s p o i n t , it i s s u f f i c i e n t t o n o t i c e t h a t in the
l i g h t o f t h i s approach The c ar rose i s non-causati ve w h i l e John rose
is vague. If John is also an agent (in the sense o f voluntarily
causing his own motion) the sentence rece ive s a caus at ive
interpretation (along w i t h an agent ive one); if not, it r ece ive s a
non-causati ve i n t e r p r e t a t i o n (along wi th the non-agenti ve one). How

- 123 -
is th is conclusion a r r i v e d at? Wi thi n t h i s system, c au s at iv e verbs
are understood as expressing a relation between two events, "the
firs t o f which i s something t h a t an agent does, the second an event
t h a t hi s a ct io n causes" ( i b i d . :475). This is p r e c i s e l y the poi nt at
which c a u s a t i v i t y and a g e n t i v i t y are c o n f l a t e d wi th what seem to me
to be unde si ra bl e results. This assumption (i.e . that t he causing
event i nv ol ve s an agent) is used in order to distinguish between
'k ill' which i mpl ie s br i ngi ng about a change in the c o n di t i o n o f i t s
logical object and ' s e e ' which does not imply any such t h i n g . The
distinction in question is related toc a u s a t i v i t y , not agentivity.
So, i nvoking the l a t t e r noti on as w e l l , in t r y i n g t o account f o r i t ,
is rather mi sl ea di ng. The fact that 'see' does not qualify as a
c a u sa t i ve has nothing to do with the intentionality or non-
intentionality o f the event. N ot ic e , at any r a t e , t h a t even i f it
i n vol ve s a v o l i t i o n a l s ubj ect in a case such as He saw the f i l m in
order to decide for h i m se l f whether the new a c t o r was an.y good,
nothing changes concerning c a u s a t i v i t y : the f i l m ' s c o n d i t i o n remains
the same.

These f a c t s are too simple t o have been overlooked by such a thorough

a n a l y s i s as the one under di scussion. The question t h e r e f o r e a rises
as to the reason why M i l l e r and Johnson-Laird (1976) opt for the
alternative approach j u s t mentioned. It seems t o be the case t h a t
neat formalization lies behind their analysis. The most economic
solution for causative/non-causative members o f pairs is certainly
the one which i ncor por at es the intransitive schema into the
transitive one and offers a s in gl e formula in the case of
phonologically identical pairs, e.g. TURN ( ( x ) , y ) . The first
argument (x) being o pt iona l in t h i s formula, the schema allows both
f o r c au sa ti ve and non-causati ve uses. Consequently, i n the case o f
phonologically non-identical pairs, a s i m i l a r method can be a p pl ie d.
Hence sentences such as He r a i s e d the car and The c ar rose are lumped
together as de sc r ib i ng "the same event" ( i b i d . :475). Since the
motions described are precisely the same in both uses, M iller and
Johnson-Laird postulate one sense of the verb. The causati ve
component is said to appear when t he re is an agent at least
i m p l i c i t l y present and t o disappear when s/he i s absent. But t h e i r
diagnostic question f o r singling out causati ves from non-causatives

- 124 -
"What p r o p e r t i e s o f y are changed by x 's a c t io n " cannot be asked in
the case of The car rose, they c la im , "because the verb is
intransitive". Now t h i s is e i t h e r f a l s e or in c o n t r a d i c t i o n to t h e i r
o t he r cl ai m t h a t i n t r a n s i t i v e s may well be c a us a t i ve s . Notice t h a t a
caus at ive understanding i s postul ated both f o r The man rose and f o r
Mel issa t r a v e l l e d although the verbs i nvolved are also i n t r a n s i t i v e .
So in r e a l i t y The c ar rose cannot be submitted to the d i a g n o s t i c t e s t
because of a prior (unadmitted) recognition of the fact that the
s ubj ect is i nanimate and t h e r e f o r e no candidate for an agent (in
their own understanding o f the term a t least). This being so, the
question o f c o r e f e r e n t i a l i t y cannot be r a i s e d as i t i s r ai se d in the
case o f The man rose and Melissa t r a v e l l e d . It is t h i s that is a t
issue and not in tra n sitiv ity as is claimed. So in essence the
difference between The car rose and The man rose is clearly a
question of agentivity (i.e . properties of the subject-causer) and
not c a u s a t i v i t y ( i . e . p r o p e r t i e s / s i t u a t i o n o f the o b j e c t of change).

The unde si ra bl e results of th is approach c o nsi st in distinguishing

between d i f f e r e n t interpretations of ' r i s e 7 (a c aus at ive and a non­
c aus ati ve one) for the sake (it seems to me) of br i ngi ng to ge t he r
'ra ise ' and 'ris e ' as c a us a t i ve s . Consider a case l i k e Smoke rose
from the chimney. There is clearly no co rresponding transitive
sentence ^Somebody r a i s e d smoke from the chimney. N e i t h e r does the
possibility of vagueness exist as it does in the case of The man
rose. Since smoke i s incapable o f "v ol unt ar y act ion " the question of
coreferentiality cannot be raised. If we adhere strictly to the
principles of M i l l e r and Johnson-Laird we can only conclude t h a t we
are faced wi th a t h i r d kind o f ' r i s e ' t h i s t i m e. This s o l u t i o n seems
more counter-intuitive than distinguishing between a causati ve
'ra is e ', and a non-causati ve 'ris e ', the relation between which
cannot rest on c on si d er a ti o ns bearing on agentivity. Different
i n t e r p r e t a t i o n s o f ' r i s e ' which depend on the s o r t o f s ubj ects i t can
accept ('the car', ' t h e man', 'smoke', etc.) need to be i n v e s t i g a t e d
s e p a r a t e l y to di sc ov e r where the verb stands as t o a g e n t i v i t y . The
relation between ' r a i s e ' and ' r i s e ' c o n s t i t u t e s an important problem
which r e q ui r es spec ia l attention but i t is a problem o f c a u s a t i v i t y
and cannot be solved by r e f e re n ce to the p r o p e r t i e s o f the causing
event and the e x t e n t to which i t i nvolves a genuine agent.

- 125 -
I have argued so far that causativity and agentivity can be
considered separately to some extent. This can be effected if
causativity is understood as c e n t r i n g around the question o f whether
something that happens (event or process) is due t o some e xt er nal
inte rvention or not, i.e. is brought about by an external cause.
Approaches attributing a caus at ive component to sentences like
Mel issa t r a v e l l e d on the grounds t h a t "she did something t h a t caused
herself to move" (M iller and Johnson-Laird 1976:544) are here
considered inappropriate as blurring the distinction between
c a u s a t i v i t y and a g e n t i v i t y . Me li ssa using her own f or ce s to t r a v e l
can be ha r dl y c a l l e d an ' e x t e r n a l ' factor, so the verb in question is
no candidate for a causative. The suggestion made here is that
c a u s a t i v i t y depends c r u c i a l l y on ' p a t i e n t i v i t y ' o f the o bj e c t r a t h e r
than a g e n t i v i t y o f the s u b j e c t , i.e. on the e f f e c t rather than the
nature of the cause. T he ref or e the degree of agentivity of the
causer i s not d i r e c t l y d e c i s i v e in c h a r a c t e r i z i n g a s i t u a t i o n as more
or less t y p i c a l l y c a u s a t i v e . Never the le ss , the noti on o f a g e n t i v i t y
plays a role in the distinction between direct and indirect
caus ati ve s in the sense t h a t the degree of p a t i e n t i v i t y o f the causee
is to some e x t en t (or in c e r t a i n cases) relatable t o the degree o f
a g e n t i v i t y o f t he causer. These claims w i l l be discussed in d e t a i l
f o l l o w i n g an account o f the r e l e v a n t noti ons: t he noti on o f ' a g e n t ' ,
the noti on o f 'tra n s itiv ity ' and the d i s t i n c t i o n between d i r e c t and
i n d i r e c t c aus at ive s.

The main idea underl yi ng a l l t h i s discussion is t h a t c a u s a t i v i t y can

be l i n k e d wi th a g e n t i v i t y through g r a d a t i o n . The assumption i s made
t h a t m a n i p u l a t i v e / d i r e c t causati on ( e . g . M a r i e t t a put the book on the
shelf) is the c e n t r a l , most c h a r a c t e r i s t i c caus ati ve situation. In
such a case the causer c o n s t i t u t e s the only f a c t o r r es po ns i bl e for
the c h a n g e - o f - l o c a t i o n o f the causee, hence the causee has no a c t i v e
role at a l l to pl ay in t he caused s i t u a t i o n , i.e. the causee is a
ty pical patient. In t h i s sense t he p a t i e n t i v i t y o f t he causee can be
considered a f un c t i o n of the agentivity of the causer. But even
direct causation does not necessarily imply that the agent is the
most ty pical one. Animacy and volition which are generally
understood as the characteristics of a genuine agent need not be
present (e.g. The earthquake p u l l e d down the house) . On the other

- 126 -
hand, indirect causation (e.g. The gaoler marched the prisoners)
r e f l e c t s s i t u a t i o n s where the causer is a t t r i b u t e d r e s p o n s i b i l i t y f o r
the caused event, yet the causee also p a r t i c i p a t e s actively in it,
i.e. s/he i s no t y p i c a l patient. Evidently, the mere f a c t t h a t the
patient is les s -th a n -ty p ic a l i n d i c a t e s t h a t the causer i s also l e s s -
than-typical; for if the caused event r e qu i r e s the active
p a r t i c i p a t i o n o f the causee as w e l l , the causer cannot be said to be
in absol ut e cont rol of the s i t u a t i o n . Yet even in the case of
indirect c aus at ion, the causer possesses (almost necessarily) the
properties of animacy ( s / he is actually human in most cases) and
v olition. Once again t h e r e is no one-to- one correspondence between
the properties of the causer and those o f the causee. This brief
o u t l i n e o f the main f a c t s o f d i r e c t and i n d i r e c t c ausati ves and t h e i r
relation to a g e n t i v i t y (which simply foreshadows a d e t a i l e d account
of these relationships) is meant to show a pos si bl e po i nt of
intersection of the notions o f agentivity and c a u s a t i v i t y and to
i n d i c a t e t h a t f o r a c l e a r e r understanding o f these problems we need a
r e c o n s i d e r a t i o n o f the noti on o f ' a g e n t ' .

The o t he r p oi n t ofintersection of agentivity and causativity is

tra n s itiv ity (as a lr ea dy mentioned). T r a n s i t i v i t y i s also understood
here as a graded phenomenon i n vo lv i ng more and l ess t y p i c a l instances
r a t h e r than a s t r i c t dichotomous d i v i s i o n between t r a n s i t i v e and non­
transitive events. In this understanding tra n s itiv ity is not a
formal syntactic cat egor y but has specific semantic content
relatable, on the one hand, to the degree of agentivity of the
s ubj ect and, on the other, to the degree of patientivity of the
object. Ther ef or e i t can only be discussed f o l l o w i n g a c l a r i f i c a t i o n
o f the notion o f agent. The e x t e n t , however, to which i t is correct
to c lai m that a categorization of c ausati ves does not necessitate
recourse to t he noti on of agentivity w ill be discussed in what
f ol l ow s i mmediately, namely an account o f MG c ausati ves o f motion.

3.2 I d e n t i f y i n g Modern Greek causati ves o f motion

Wi thi n the area under investigation it seems re la tiv ely easy to

identify those verbs which are causati ves provided one accepts the

- 127 -
traditional link between causati ves and transitives, i.e. the
requirement t h a t for something t o q u a l i f y as a c au sa ti ve i t must a t
least be a transitive verb. As a lr e ad y noted in the previous
section, causati ves of motion are here understood as d e sc r ib i ng a
change-of-location/position explicitly attributed to ext er na l
intervention. The prerequisite can be therefore s ta t ed in very
simple terms: caus at ive s o f motion are expected t o gi ve rise to a
c o n t r a d i c t i o n i f they appear in a sentence o f the form ' x Ved y but y
did not move'. Three main poi nts have t o be discussed a t t h i s stage:

whether the l e x i c a l i z a t i o n o f the r e s u l t i n g state (position) of

the o bj e c t which is caused to move plays a role in identifying
causati ves o f motion;

whether t he re i s any evidence f o r c lai mi ng t h a t t he i n t r a n s i t i v e

( no n- c au s at i ve ) or the causati ve i t corresponds t o i s more basic
(in those cases obvi ousl y where such a correspondence can be
established), i . e . whether t h er e i s reason to b e l i e v e t h a t one is
de ri ve d from the o t h e r ;

whether we can t a l k about r e l a t i v e prominence o f the causati ve

element in c e r t a i n verbs ( or sets o f verbs) and i f so what is the
nature o f the phenomenon.

As was poi nted out in the previous section, Lyons (1977) suggests
t h a t i f a verb i m pl ie s t h a t an e f f e c t i s unequi voc al ly brought about,
i.e. if it is a clear case o f a caus ati ve (in my under st anding) ,
this effect must lexicalize. The v a l i d i t y of this obser vati on for
all causati ves implying a c hang e- o f- c on di ti on o f t h e i r o b j e c t cannot
be tested here. It is worth, however, examining whether such a
requirement is at all necessary in the case o f the MGMVs which are
here t e n t a t i v e l y posi ted as causati ves o f motion.

Four groups o f t r a n s i t i v e MGMVs can be d i s t i n g u i s h e d on the basis of

whether and how the change-of-locatio n/position of the obj ect
lexicalizes. Those most easily identifiable as caus at ive s are
perhaps the ones which happen to have a phonologically identical

- 128 -
intransitive counterpart. In the field under c o ns id e r a t i o n the
category i ncludes verbs such as:

I. 'kilao2 ' ( r o l 1*j-r) , ' plisiazog' (approachyp), ' strivoz' (turnyp),

lyirizo 2/ (turnyp), ' v u l j a z o 2' (drownyp), 'akubao2' (make s . t h .
touch s.wh., put), '\ ona tiz 0 2 ' (make s.o. kneel), ’ xami
(lowerjp), ' epistrefo2 ' (returnyp), 'yirizo pi s o 2 > (returnjp),
' y l i s t r a o 2/ t s u l a o 2 r (slipyp), ' p/ yeno^ ' (take s.th. s.wh.),
'paramerizo2l (pull, put aside), ,kikloforo 2' (put into
c irculation/circulateyp).

I t must be poi nted out t h a t t h i s group i ncludes verbs which are very
different semantically, accepting distinct kinds of causers and
causees and d e s c r i bi n g a v a r i e t y o f motions. There i s no u n i f o r m i t y
as t o the exact contribution of the causer/causing event. So the
p r i v i l e g e d s tatus such 'double purpose' verbs are sometimes accorded
in analyses o f English c a u s a t i v e / n o n - c a u s a t i v e p a i r s o f verbs cannot
be accepted f o r MGMVs, a t l e a s t , on the basis o f semantic c r i t e r i a .
The reasons o f f e r e d f o r English 'double purpose' verbs are not very
clear-cut e ith e r. As a f a i r l y s t r a i g h t f o r w a r d example, consider the
sort of sentences used in accounts of the character of sentence
p r on omi n a l i z a t i o n :

(1) John f i n a l l y melted the metal but i t took him years to bri ng
i t about.

(2) *John f i n a l l y k i l l e d Mary but i t took him years t o bri ng i t

about.

I gnori ng interpretations o f the type John t r i e d over and over again

as irrelevant, the f a c t remains t h a t : (1) is acceptabl e and (2) is
not. L a ko f f and Ross (1972) attempt an e xp l an a ti o n o f these f a c t s
along the f o l l o w i n g l i n e s : the a c c e p t a b i l i t y o f ( 1 ) is a ttrib u te d to
t he morphological relation between caus ati ve and intransitive verb
( e . g . harden, melt e t c . ) , w hi le the u n a c c e p t a b i l i t y o f (2) i s claimed
to be due to the f a c t t h a t "the l e x i c a l item and the antecedent are
not mor p ho l og i ca l ly r e l a t e d " ( p . 122). Not ic e, however, that (3) is

- 129 -
acceptabl e while (4) is not , although the opposite fa cts hold in
connection wi th ' phonol ogi cal iden ticity':

(3) John f i n a l l y t ra n s p o r t e d t he trunks t o Afgani stan but i t took

him years t o br ing i t about.

(4) ??John f i n a l l y turned the switch l e f t but i t took him years

to bri ng i t about.

In short, even in English such phenomena are probably better

accounted for in terms o f the 'processual' or 'event' c h ar a c t e r of
the predications in question rather than identity of form between
c a us a ti ve and non-causati ve members of verb p a i r s . Be t h a t as it
may, no speci al semantic c h a r a c t e r i s t i c s can be a t t r i b u t e d to group I
MGMVs. All t h a t can be s a f e l y said about them is t h a t t h e i r meaning
can be most d i r e c t l y computable on the basis o f the meaning of t h e i r
intransitive c ount er par t s plus a c au sa ti ve element (in its broadest
under st and ing) . In the case o f MG t h i s fact is not devoid o f any
seri ous implication. There is fa irly strong evidence that such
c au sati ves are d e r iv e d from t h e i r i n t r a n s i t i v e c ount e rpa rt s which can
therefore be considered more basic (in this respect). In recent
years a g r e at number of new causati ves have emerged (especially in
the speech of the younger g e ne r a t i o n ) which correspond to a lr eady
existingin tra n s itiv e s . So, parallel to colloquial expressions in
c u r r e n t use f o r many y e a rs , e.g . ' t o n peQane' ( s/ he 'died' him, i.e.
s/he made him die), new expressions are now used, e.g. 'na to

k a t a l i k s ut ne to Qema' ( l e t us 'conc lud e' the i ssue, i.e. l e t us bring

the issue to an end and reach a c o n cl u s i o n ) . This expression was
heard w i t h some s u r p r i s e two years ago in a s t u d e n t s - s t a f f meeting at
Athens U n i v e r s i t y but was immediately i n t e r p r e t e d i n the way i t was
meant to be and has by now become q u i t e 's tandardized'. In s h o r t ,
there is reason to b e l i e v e t h a t in t he case o f 'double purpose' MG Vs
there is a process i nvolved in the d e r i v a t i o n o f t he causati ves and
that this process is, moreover, quite p r od u c t i v e . So, in this
p a r t i c u l a r case i t makes sense t o p o s i t the i n t r a n s i t i v e as the basic
form and mark in the l ex i co n those intransitives which undergo the
process of being used also as c aus a ti ve s . More evidence in this

- 130 -
direction w ill be discussed a f t e r the p r es en t a t i o n o f t e s t r e s u l t s in
the f i n a l c hapt er .

The second group o f MG causati ves o f motion, s ingled out again on

morphological rather than semantic c riteria, i nvolves verbs which
have a mor ph ol o gi ca l ly r e l a t e d (although not i d e n t i c a l ) intransitive
counterpart. Such verbs are:

II. 'anevazo' (take up) corresponding to 'aneveno' (go up, ascend),

'katevazo' ( t ak e down) corresponding to 'kateveno' (go down,
descend), ' bazo' (put in) corresponding to ’ beno’ (enter) and
' vyazo' ( t ak e o ut ) corresponding to ' v y e n o ' (go o u t ) .

Once again the correspondences are q u i t e easy to e s t a b l i s h ; (5a) can

be said to imply ( 5 b ) :

(5a) anevasa t a vivlia sto rafi (V=anevazo)

I took-up the books on the s h e l f
I put the books up on the s h e l f

(5b) ta vivlia anevikan sto rafi (V=aneveno)

the books went-up on the s h e l f
the books went upon to the s h e l f .

Besides the lexicalization of the event w it h a non-causati ve

( 'a n e v e n o '), the new p o s i t i o n can also l e x i c a l i z e wi th a p r e d i c a t i o n
which i s as s t a t i v e as i s compatible wi th the noti on of motion, i.e.:

(5c) ta vivlia ine anev as men a s t o rafi ( V = a n e v a z o me )

the books are taken-up to / on the s h e l f
the books are put up on the s h e l f .

It would seem a t firs t sight that since these verbs are closely
related in meaning, the l e x i c a l i z a t i o n o f the c h a n g e - o f - l o c a t i o n is
in t h e i r case l es s a r b i t r a r y than f o r ot he r verbs. Notice, however,
that two more verbs wi th very similar morphological and semantic

- 131 -
c h a r a c t e r i s t i c s can be added to t h i s group, f o r which no i n t r a n s i t i v e
c o u n te r pa rt exists (with the same p r o p e r t i e s as those o f the verbs
presented under I I ) , such as ' v a z o ' ( put ) and fbizo' (stick (in)).
The r e s u l t i n g position o f the o bj e ct can l e x i c a l i z e in the case o f
'vazo' w i t h i n a s t a t e p r e d i c a t i o n which makes no r e f e r e n c e whatsoever
to the pr ecedi ng/ causi ng event. This i s e f f e c t e d through monovalent
predicators which have no c ausati ve c o un t e r p a r t and are
mor phol ogi ca ll y compl etel y u nr e la te d t o ' v a z o ' , namely ' i m e ’ (be) and
'vriskome' (be f ound), e . g . (6a) can only ' r e s u l t ' in ( 6 b ) :

(6a) evala ta vivlia sto trapezi (V=vazo)

I put the books on the t a b l e

(6b) ta vivlia ine/vriskonde sto trapezi ( \ = i m e / v r i skome)

the books are (found) on the t a b l e

In those cases where 'vazo' ( put ) is fol lowed by a PP i n v o lv i n g

'inside something' the verb can 'borrow' from 'bazo' (put in) a
monovalent p r e d i c a t o r r e f e r r i n g to the event o f the o b j e c t ' s change-
o f-location, e.g. (7a) i mpl ies ( 7b) :

(7a) evala to vivlio sto sirtari ( \i=vazo)

I put the book in the drawer

(7b) to vivlio bike sto sirtari ( V= b e n o )

t he book w e n t - i n the drawer

This can be i n t e r p r e t e d as implying t h a t the reason why ' v a z o ' (put)

does not have a l e x i c a l i z e d monovalent p r e d i c a t o r as ' a n e v a z o ' (take
up) and the ot he r verbs o f group I I have, i s p r e c i s e l y the semantic
fact that, as a cover term o f those verbs, it does not s p ec i fy how
the motion i s e f f e c t e d w it h r e l a t i o n t o the r e s u l t i n g p o s i t i o n o f the
moving o b j e c t . This p o s s i b i l i t y cannot be r ul e d out. It is of no
great importance, however, because 'b izo ' (stick (in)) which is at
least as s p e c i f i c as ’ bazo' (put i n) does not have an i n t r a n s i t i v e
c o u n te r pa rt sim ilar to those of group II verbs either. The

- 132 -
conclusion t h a t l e x i c a l i z a t i o n o f the c h a n g e - o f - l o c a t i o n / p o s i t i o n is
again arbitrary seems i nescapabl e. In any case, the causati ve
c h ar a c t e r of both 'vazo' and 'b izo' (lacking a corresponding
monovalent predicator) cannot be di sputed, and neither can their
close morphological and semantic relationship wi th the causati ve
verbs o f group I I . Yet no uniform e xpl ana ti on can be found f o r the
presence (or absence) in this subfield of a monovalent predicator
lexicalizing the c h a n g e - o f - l o c a t i o n . Besides, t h e re seems to be no
ground f o r e s t a b l i s h i n g the in transitive r a t h e r than the t r a n s i t i v e
as more basic in the case of these verbs; so we can only assume t h a t
the question o f d e r i v i n g one from the other cannot be r a i s e d .

A third category of MG causati ves of motion can be identified on

morphological grounds. It comprises verbs which may be said to have
a corresponding monovalent p r e d i c a t o r which i s mediopassive in form,
i.e. ending in ' - o m e ' . It is easy to f i n d examples o f t h i s category
but very d iffic u lt to decide which verbs are genuine members and
which are not, the reason being that for a large number o f '-ome'

verbs o f motion, at le a s t, it i s not in the l e a s t c l e a r whether they

are passive or not. The whole issue is of great importance for
c a u s a t i v i t y and a g e n t i v i t y . Ther ef or e ' - o m e ' verbs are given special
attention i n a subsequent s e c t i o n . For the moment, a few verbs w i l l
be presented as examples o f what I consider group I I I here in order
to see how they compare to verbs in o t he r groups, always in r e l a t i o n
to the questions o f l e x i c a l i z a t i o n and d e r i v a t i o n .

III. 'sikono' (raise) corresponding to 'sikonome' (rise),

' apomakrino' (remove, move awayyp) corresponding' to
' apomakrinome' (move awayiNTR)> 'tinazo' (shake upyp)
corresponding to 'tinazome' (' shake up'j^yp, be shaken up),
' v i 07z o ' (drowny^, s i n k, immerse in w at e r ) corresponding to
'viQizome' (drown^y^, sink), 'gremizo' (pull down)
corresponding t o ' gremizome' (fall down (a p r e c i p i c e ) ) , 'xono'
(stick(in )), corresponding to 'xonome' (be s t uc k/ e ngul fe d i n ) ,
'kaOizo' (make s.o. sit) corresponding to ' kaQomei' (sit, be
seated), etc.

- 133 -
The semantic r e l a t i o n o f verbs in t h i s group to those o f o t h e r groups
is rather evident. Notice, for i nstance, the relation of 'sikono'

(raise) to 'anevazo' ( t ak e up), ' xono' (stick (in)) to 'bizo'

(s tic k (in )), 'viB izo' (sink, drown) to 'vuljazo' (sink, drown),
'kadizo' (make s.o. s it) to 'y o n a t i z o 2 t (make s.o. kneel ) and
' apomakrino1 (move away) to ' paramerizo' (pull, put aside). They
d iffer from one another in the sorts of causers and causees they
accept and the exact involvement of the causer. Notice, however,
that such differences appear both within each group and across
groups, so once again the type o f l e x i c a l i z a t i o n o f the change-of-
l o c a t i o n does not appear t o have any semantic s i g n i f i c a n c e . Besides,
whether they have a more or l ess indisputably intransitive '-ome'

c o un te r pa rt (e .g. 'sikono' (raise)) or are only r e l a t e d to a passive

'-ome' verb (which does not normally count as ' l e x i c a l i z e d e f f e c t ' ) ,
they always e n t a i l t h a t t h e i r d i r e c t o b j e ct changes l o c a t i o n owing to
e x t er n al intervention and therefore pass the test for c aus a ti ve s .
F inally, although '-ome' i n t r a n s i t i v e s can be e t y m o l o g i c a l l y analysed
as 're fle x iv e s ', there is no evidence f o r p o s t u l a t i n g the c ausati ve
forms as more b a si c, S yn c hr o n i c a l l y , the process of
're fle x iv iza tio n ' is productively effected through completely
different means ( i . e . addition o f the p r e f i x 'a fto -') and does not
concern the verbs in quest ion.

I V. The fourth group of verbs is understood here as i n v ol v i ng

causati ves which either have a count er par t completely u n r e l at e d to
them in form, or no c ount e rpa rt a t a l l . This i s a r a t h e r d i f f e r e n t
case from those discussed so f a r . In t he absence o f morphological
relations it is not equally easy to tell what t he new state of
affairs ( t he changed location/position) actually involves. It is
therefore worth investigating: (a) whether s p e c i f i c correspondences
can be e s t a b l i s h e d f o r some verbs and (b) whether absence o f such a
correspondence affects the c ausati ve character of the verbs in
quest ion.

Certain correspondences are fa irly easy to establish, A case in

p oi nt i s the p a i r 'ferno' (bring) - 'erxome' (come). The s i m i l a r i t y
o f t h i s p a i r to t p i \ i e n c 2 t ( t ak e s . t h . s .w h. ) - 'piyenoj' (go) is too
obvious to r e q u i r e e l a b o r a t i o n . Other cases such as t h a t o f ' s t e l n o '

- 134 -
(send) and ' p e t a o / r i x n o ' (throw) are more pr obl emati c and should be
considered in some d e t a i l .

In certain examples 'stelno' (send) may be said to entail 'ftan o '

( a r r i v e ) , w hi le in others t h i s correspondence i s not v a l i d :

(8a) *estile ti bala sta Sixtja, ala i bala 5 en e f t a s e e k i

s/he sent the b a l l t o the g o a l - p o s t , but the b a l l did
not a r r i v e / g e t t he re

(8b) estile ena y r a ma s t i yermania, ala t o y r a ma 5 en e f t a s e

(eki)

s/he sent a l e t t e r to Germany, but the l e t t e r di d not

get/arrive (there).

Appar entl y what 'ste ln o ' necessarily i mpl ies is t h a t the o bj e c t is

caused t o start travelling (i.e . l eave the pl ace i t used t o hold)
hence the u n a c c e p t a b i l i t y o f ( 8 c ) :

(8c) *estila ena y r a ma s t i yermania ala ine a k o mi sto sirtari mu

I sent a l e t t e r to Germany but i t is s t i l l i n my drawer

What it does not imply is that the goal intended by the causer is
reached, hence the acceptability of (8 b). The unacceptability of
(8a) can be expl ai ned i f one considers t h a t the j our ney i nvolved is
so short t h a t the observer would be expected to take i n the whole of
it r a t h e r than j u s t the beginning. In t h i s re spe ct 'stelno' (send)
is not u n l i k e verbs such as ' p e t a o 2l (throw) or 'ektoksevo' (hurl)
which have a sim ilar restriction. The o bj e c t is caused to travel
(through the air in these latter cases) but its resulting position
w ill only l e x i c a l i z e under c e r t a i n conditions. The natur e o f these
condi ti ons has to be examined in order to see whether it can be
r e l a t e d t o the issue o f r e l a t i v e sal ience/ promi nence o f the causati ve
element. S i m i l a r l y to ' s t e l n o ' (send), fpetao2f (t hrow) and ' r i x n o '
(throw) may be seen as partially entailing 'pefto' (fa ll), on the
basis o f examples such as ( 9 a ) :

- 135 -
(9a) *petakse/erikse to kuti s t o patoma, ala to k u t i ben epese
s t o p a t o ma

s/he threw the box to the f l o o r , but the box di d not f a l l

on the f l o o r .

It might seem a t firs t sight that the crucial difference between

' t h ro w ' and ' f a l l ' i s t h a t the l a t t e r necessarily i nv ol ve s downward
motion u n l i k e the former, hence the correspondence i s restricted to
those cases where t h i s c o ndi ti on is also met by 'throw'. No t ic e ,
however, that although t h i s is t r u e to some e x t e n t , a specification
o f the goal (reached and maintained by the moving o b j e c t ) i s perhaps
equally i mportant w i t h respect t o the f a c t s o f MG ' p e t a o / r i x n o ' and
'pefto' at le a s t. So (9b) i s also unacceptable f o r the same reason
t h a t (9a) i s , although the motion described i s not downward:

(9b) *erikse to aftokinito s to bendro, ala to aftokinito ben

epese s t o bendro

s/he ' t h r e w ' the car on the t r e e , but the c ar di d not

fall on the t r e e
s/he crashed the car i n t o the t r e e , but the c ar di d not
h i t the t r e e .

I f t he d i r e c t i o n o f the moving o b j e c t is upward, e . g . :

(9c) erikse ti bala p s i l a / s t i steyi

s/he threw the b a l l h i g h / t o the r o o f

the condi ti ons of 'goal specification' and 'downward d i r e c t i o n '

c oi nci de i n the sense t h a t g r a v i t y w i l l also determine the new and
final position of the moving o b j e c t . Ther ef or e in (9c) 'p sila '
( hi gh) w i l l not a ll o w a l e x i c a l i z a t i o n wi th ' p e f t o ' ( f a l l ) w h i le ' s t i
steyi' ( t o the r o o f ) w i l l .

The facts of English ' t h r o w ' are not dissim ilar t o those just
presented f o r ' p e t a o £ ’ and ' r i x n o ' , so i t i s worth l ooking c l o s e l y at
an a n al y s i s o f ' t h r o w ' which draws a l i n e between (10a) and (10b) in

- 136 -
terms o f the r e l a t i v e prominence of the feature o f motion (Ikegami
1 96 9: 89) :

(10a) He threw the b a l l s k illfu lly

(10b) He threw the b a l l over the fence.

The former sentence is analysed as "make a c e r t a i n movement of the

arm + in order to cause something to go through the air" and the
latter as "cause something to go through t he air" + "by making a
c e r t a i n movement o f the arm". I kegami ' s view is t h a t the d i f f e r e n c e
between uses such as t h a t o f (10a) and the one i n (10b) lies in the
relative prominence of "the feature of motion" in connection wi th
t h i s verb. Hence he c at e g o r i z e s such verbs as being " s i t u a t e d on the
fringe o f the category o f verbs of motion" (p .90). It seems to me
t h a t the d i f f e r e n c e between the two examples o f f e r e d by Ikegami might
become clearer if a third common use of 'throw' is juxtaposed to
them; in p a r t i c u l a r , one which s p e c i f i e s the goal o f t he motion even
more p r e c i s e l y than (10b) and can imply t h a t the new p o s i t i o n (goal
reached) is preserved. Consider t h e r e f o r e the example:

(11) He threw the b a l l t o the t r e e .

This, I suggest, may be more d i r e c t l y understood as i mpl yi ng (among

ot he r t h i ng s ) 'he caused the b a l l to f a l l o n / a t the t r e e ' and allows
f o r the p o s s i b i l i t y t h a t the o b j e c t stayed a t i t s new l o c a t i o n . Now
(10b) may be also understood as implying 'he caused t he b al l to f a l l
a t a place which was a t the ot he r side o f the f e n c e ' . Sentence (10a)
is s t i l l one step f u r t h e r than (10b) in t h a t the goal o f the o b j e c t ' s
jour ney is compl etel y unspecified. This seems to be the crucial
difference wi th t he remaining examples. To what extent is it
therefore true to say that for ' t h ro w ' "the non-causati ve
interpretation is t he more usual one" as Ikegami does? (1969:105).
On t he basis o f the examples discussed i t seems t o be the case t h a t
two i m p l i c a t i o n s remain constant: (a) the o b j e ct i s caused t o t r a v e l
through the a i r (b) i t r ec ei ve s impetus by the causer. The c o ndi ti on
o f motion i s always prominent ( c o n t r a r y to I kegami 's c lai ms) and the

- 137 -
condition of causativity is always p re se nt . If the goal is not
s p e c i f i e d or not reached the r e s u l t i n g l o c a t i o n o f the o bj e c t is not
registe red and the c h a n g e - o f - l o c a t i o n does not l e x i c a l i z e through a
specific intransitive. This does not render the c aus at ive element
l ess prominent. On such i n t u i t i v e grounds the exact opposi te could
also be clai med, namely t h a t the c o n t r i b u t i o n of the causer is more
crucial; in a way, all we know about the event in question is t h a t
some o b j e c t is caused t o travel by some e x t er n a l source o f motion.
I t seems, however, t h a t the r e l a t i v e prominence o f the causer/causing
event cannot be discussed on the basis of such c o n si d e r at i o ns .
Factors such as p e r c e i v a b i l i t y o f the causer, degree o f involvement
of the causer (in the pr ocess/ event caused), accompaniment, e tc.,
seem to be much more r e l e v a n t than presence or absence o f l e x i c a l i z e d
effect. As a lr e a d y poi nted out , such factors, which bear on the
na tur e o f the causer r a t h e r than the e f f e c t , are best discussed in
connection wi th agentivity and will be taken up later. For the
moment, all t h a t can be said on the basis o f the cases discussed so
f a r i s t h a t t he re i s no evidence:

t h a t the l e x i c a l i z a t i o n o f the c h a n g e - o f - l o c a t i o n / p o s i t i o n of the

o b j e c t plays a r o l e in i d e n t i f y i n g causati ves o f motion;

t h a t e i t h e r the c au s at iv e or the non-causati ve verb is more basic

except in the case o f p ho no l o g i c a l l y i d e n t i c a l p a i r s .

On the o the r hand, relative prominence of the caus at ive element

r e q ui re s recourse to the noti on of agentivity. This is therefore
posi ted as a p o i n t o f i n t e r s e c t i o n o f the notions o f c a u s a t i v i t y and
agentivity.

3.2.1 Non-causative t r a n s i t i v e s

In the s p e c i f i c area under i n v e s t i g a t i o n t h e r e are two c a t eg or i es of

transitives which are not c aus a ti ve s . Both c a t eg or i e s describe a
s i t u a t i o n where an e n t i t y x moves in r e l a t i o n to some o t he r e n t i t y y

- 138 -
which is either not moving {usually expressed through a NP of
l o c a t i o n ) or moving but not n e c e s s a r i l y a t the i n s t i g a t i o n o f x .

The f i r s t one comprises cases such as:

I. (a) 'p lisia zo tin poli' (approach the c i t y )

(b) 'biasxizo tin p latia' (cross the square)
(c) 'bjaveno ton potamo' (cross the r i v e r )
(d) 'pernao t o bromo' (cross the road)
(e) ' prospernao to mayazi' (pass (by) the shop)
(f) ' s k a r f a l o n o t o vuno' (cl imb the mountain)
(g) 'strivo ti yonia' (turn ( a t / r o u n d ) the corner)
(h) 'pibao to fra xti' (jump (over) the fence)
(i) ' aneveno t i skala' (go up the l adder )
(j) ' kateveno to potami' (go down the r i v e r )
(k) ' p e r p a t a o mj a a p o s t a s i / 2 0 x i l j o m e t r a ' (walk a distance/20km)
(1) ' t r e x o mja a p o s t a s i / 2 0 x i l j o m e t r a ' (run a di stance/20km)
(m) ' k o l i b a o mja apostasi/20 x iljo m e tra ' (swim a distance/20km)
(n) 'taksibevo t o n k o s mo ' (travel (around) the w o r l d ) .

These can be regarded as intransitive verbs which under certain

circumstances - such as the ones e xe mp l if ie d above - become
transitives w it h an NP o f location as their direct object. They
cannot imply t h a t t h i s o bj e c t i s caused t o move; only the subj ect is
moving. Notice, however, that ' p l i s i a z o 2' (approachjp) can be a
caus at ive in cases like /p l i s i a z o tin karekla sto trapezi' (I
' approach' the chair to the t a b l e , i . e . ' I move the c h a i r near the
table). Rather predictably t h er e are no passive const ruct ions
equivalent to examples (a) to (n) since the o b j e ct NPs are not
lYlOX
a f f e c t e d by the a ct i o n o f the s u b j e c t . This doesAmean, however, t h a t
there are no mediopassive forms corresponding to the verbs in
que st ion. It is i mportant to notice, though, that under certain
circumstances, for some o f these verbs, passive c o nst r uct io ns are
possible, although the verbs are s till non-causati ves of motion,
e.g.:

- 139 -
(e') / vo 1 vo p r o s p e r a s t i k e apo ena f i a t (V=prospernjeme)
the Volvo was overtaken by a F i a t .

Notice also t h a t although (12a) does not have a corresponding (12b):

(12a) o janis etrekse eksi m ilja (V=trexo)

John ran s i x mil es

(12b) *trextikan eksi m ilja apo t o n jani ( V=trexome)

s i x miles were run by John

the mediopassive form o f the same verb (i.e . 'trexome') is possi bl e

in cases l i k e the one e xe m p l i f i ed in ( 1 3 ) :

(13) trexete afti i apostasi?

i s - i t - r u n t h i s distance?
i s i t possi bl e to run t h i s distance?

On the basis of such data it becomes obvious that the relations

between c onst ruct ions such as (13), transitive c ausati ve
c o n st r u ct i o n s, and t r a n s i t i v e non-causati ve ones have t o be examined
more c a r e f u l l y , and t h a t passivizability can be used as a tool in
this investigation. Factors such as 'degree o f affectedness o f the
o bj e c t of a transitive construction' play an i mportant role in
whether a c e r t a i n t r a n s i t i v e verb can have a corresponding passive in
MG. Presence or absence o f c a u s a t i v i t y cannot solve such problems by
i t s e l f and although in most cases i t i s easy t o determine ( w i t h i n the
field under investigation) whether a verb is caus at ive or not,
marginal cases also e x i s t as w i l l become obvious in what f o l l o w s .

The second category o f non-causati ve t r a n s i t i v e s o f motion comprises

verbs such as:

II. 'kiniyao' (hunt), 'akoluQo' (follow ), ' k a t a b i o k o ' (c ha se ),

' sinobevo' (accompany), 'obiyo' (lead).

- 140 -
In a sentence 1 ike (14):

(14) to pebi sinobefse tus proskopus s t i n korifi (V=sinobevo)

tu lofu

the c h i l d accompanied the boy scouts t o the top

o f the h i l l

it cannot be claimed that the s ubj ect is the causer o f t he direct

object's change-of-location. There is no i n d i c a t i o n that the event
would not have taken place w i t h ou t the subject's intervention. A
case l i k e (15) seems, however, more pr obl emati c:

(15) 7 astinomia katabioke ton klefti apo x o r j o (V=katabioko)

se x o r j o

the p o l i c e chased the t h i e f from v i l l a g e

to v i l l a g e

Probably the most immediate i n t e r p r e t a t i o n o f t h i s sentence would be

t h a t the t h i e f ' s continuous c h a n g e - o f - l o c a t i o n ( ' f r o m one v i l l a g e to
the n e x t ' ) i s due t o the chasing o f the p o l i c e : if the p o l i c e were
not after him, the thief might not have moved at a ll. Not ice,
however, t h a t the c h a n g e - o f - l o c a t i o n described by the PP in question
does not necessarily refer to the th ie f's movements but rather to
those o f the p o l i c e . Hence (15) is by no means an unnatural sentence
(although i t does not i nvolve the most p r o t o t y p i c a l understanding o f
the verb i n q u e s t i o n ) :

(16) 7 astinomia katabioke ton klefti pu (\= k a ta b io k o )

krivotan s e mj a s p i l j a

the p o l i c e were chasing the t h i e f who

was hi di ng in a cave.

N ot ic e , f u r t h e r , t h a t the o bj e c t of chasing need not even be aware o f

the fact that s/he i s being chased. It can be t h e r e f o r e said t h a t
the verb i s mainly d e s c r i p t i v e o f the s u b j e c t ' s motion and e s p e c i a l l y
of his/her intentions. In this sense i t is not considered here a

- 141 -
caus ati ve o f motion, although the p o s s i b i l i t y t h a t i t may f u n c t i o n as
one is not r u l ed out .

The least clear case in category II is actually 'obiyo' (lead).

Consider sentence ( 1 7 a ) :

(17a) t o p e b i me o b i y i s e sto sp iti tu ja n i (V=obiyo)

the c h i l d l ed me to John's house.

This sentence n e c e s s a r i l y e n t a i l s :

(17b) p i y a / e ftasa s t o sp iti tu ja n i ( \J=piyeno)

I went t o / a r r i v e d a t John's house.

It seems once again that the verb describes mainly the subject's
action, although the o b j e c t ' s motion may well be a t t r i b u t e d to t h i s
action. Consider f o r i nstance ( 1 8 ) :

(18) o xorikos obiyise to aloyo sto x o r a f i (V=obiyo)

the peasant led the horse t o the f i e l d .

Sentence (18) seems q u i t e s i m i l a r t o (19) c ont ai ni ng an indisputably

caus ati ve verb:

(19) o xorikos piye to aloyo sto x o r a f i (\l=piyeno)

the peasant took the horse to the f i e l d .

The main d i f f e r e n c e between the two l a s t sentences i s t h a t the former

s p e c i f i e s q u i t e e x p l i c i t l y t h a t the animal was using i t s sel f- movi ng
mechanism t o change l o c a t i o n i n the d i r e c t i o n i n d i c a t e d by the person
l ea di ng i t , w h i l e the l a t t e r may also be understood as implying t h a t
the horse was c a r r i e d to the f i e l d on something e ls e which was moving
(e.g. a vehicle). Hence although 'obiyo' (lead) is not immediately
understood as 'causi ng somebody's m o t i o n ' , t h i s p o s s i b i l i t y cannot be
rul ed out either. The main consi de ra ti ons that go into such a

- 142 -
d eci si on verge on who i s r espons ibl e f o r the motion, i.e. who has the
in itiativ e for t he object's motion. Such properties I regard as
mainly linked wi th the issue of agentivity. I therefore consider
t h a t a proper understanding o f an in between case l i k e 'obiyo' and o f
d i r e c t versus i n d i r e c t causati ves r equi re s a b e t t e r understanding of
agentivity.

3.3 A g e n t i v i t y as a c l u s t e r o f p ro p e r t i e s

Surveys of work on a g e n t i v i t y (e.g. Cruse 1973, Morley 1983) have

tended to concentr ate on d i f f e r e n c e s between the var ious accounts,
thereby implying that what they have in common is unproblematic.
However, f o r the purposes o f the present a n a l y s i s , at l e a s t , certain
o f the g e n e r a l l y accepted views need to be reviewed.

Fillmore ( 1968b:24) def ine s the agenti ve case as that of the

"typically animate percei ved i n s t i g a t o r o f the a c t io n identified by
the ver b". Gruber (1976:165) identifies a gent ive verbs as those
"necessitating an Animate w i l l f u l subject". Ross (1972) and Dowty
(1972b) have a ct io n verbs involving two-pl ace p r e d i ca t e s where the
firs t argument is an agent c h a r a c t e r i z e d by i n t e n t i o n (Ross) and by
intention and v o l i t i o n (Dowty). The same p o i n t is in essence taken
up by Jackendoff (1976) where the agent f i l l s up the f i r s t argument
position of a volitional predicate CAUSE. The agreement on this
issue o f proponents o f w i de ly d i f f e r e n t t h e o r i e s is r a t h e r s t r i k i n g .
The tests typically used to identify agents or agent ive verbs
involve, equally commonly, manner a d ve r bi a ls specifying 'purpose'
such as 'carefu lly', 'de lib era te ly', 'eagerly', 'a tte n tiv e ly ' which
are r e f e r r e d t o here as ' pur posi ve Advs'. Besides these, the DO-SO
test is used for Engl ish, which need not be discussed as it is
c o nv inc i ngl y proven by Langacker (1975) to be r a t h e r c o n t r o v e r s i a l .

The consequence o f such a narrow 'yes or no' understanding o f the

notions o f 'agent' and ' a g e n t i v e verb' seems to be lumping to ge t he r
verbs which can be shown t o d i f f e r s i g n i f i c a n t l y w i t h r espect t o t h i s
property. Since we are only concerned wi th motion verbs here, Gruber

- 143 -
(1976) and M i l l e r and J ohns on-La ird' s (1976) examples w i l l be given
speci al attention.

Gruber's examples o f cases where the theme can be i d e n t i f i e d as agent

i n vol ve :

(20) John went i n t o the room

(21) John r o l l e d down the h i l l

(22) John f l o a t e d across the l a k e .

These are juxtaposed to (23), meant as an illu stratio n of the

im possibility o f having an agent in the case o f inanimate e n t i t i e s
(as s u b j e c t s ) :

(23) The log f l o a t e d across the l a k e .

This is fol l owed by the obser vati on t h a t there are "very few verbs
which are Motional or Dur at iona l and which cannot be i n t e r p r e t e d as
being Agentive when the s ubj ect is Animate" (1976:158). There is
nothing wrong w it h these observations except t h a t they are not very
revealing. A closer l ook at the very examples provided ((20) to
( 2 2 ) ) shows t h a t t h e i r subj ects do not bear e x a c t l y the same r e l a t i o n
to t h e i r r e s p e c t i v e p r e d i ca t e s concerning the pr ope rt y in questi on,
even if it is restricted to intentional i t y as the onl y d e ci s iv e
factor. Purposive c ons t ru ct io n s are both uncommon and unnatural in
combination wi th (21) and r a t h e r u n l i k e l y in combination wi th (22),
w hi le (20) is (by its elf) more likely than not to r e ce i ve an
'intentio nal' interpretation. Alternatively, w h i le (21b) and (22b)
seem a l l right and the l a t t e r one i s r emi ni sc ent o f (23), (20b) is
r a t h e r strange:

(21b) John r o l l e d down the h i l l unconscious

- 144 -
(22b) John f l o a t e d across the l ak e unconscious

(20b) John went i n t o the room unconscious.

Thus the criteria of 'animacy' and ' intentionality' cannot be

sufficient. Gruber does, however, i d e n t i f y what we might c a l l here
c e n t r a l - c o r e or p r o t o t y p i c a l l y agent ive verbs as those whose subjects
are o bligatorily rather than optionally agents, e.g. 'run' and
'fle e '. The remaining English motion verbs presumably optionally
take agents as s u bj e c t s , though; therefore no f u r t h e r distinctions
are deemed necessary.

A rough but more detailed classification of verbs in terms of

agentivity is attempted by M iller and Johnson-Laird ( 1 9 7 6 ) . Some
verbs are said to i nv ol ve t h i s notion in the sense t h a t the question
can a r i s e as to whether the act was i n t e n t i o n a l or a c c i d e n t a l ; e.g.
'k ill' in a sentence like He killed a man may be understood as
intentional, and in that case it is a gent i ve ; or u n i n t e n t i o n a l , in
which case it is non-a ge nt iv e. Another category o f verbs is
e xe m p l i f i ed by 'die' and t h e r e the question o f a g e n t i v i t y does not
arise at all ( t hey could be understood as non-agenti ve par
excellence). In addition to these two, one more category is
recogni zed, where i n t e n t i o n a l i t y is incorporated i n t o the meaning o f
the verbs in q uest ion. In such cases the question o f agentivity
again does not arise, within this system, since it is a necessary
condition, e.g. 'chase' in They have been chasing him since dawn.
These verbs could be t h e r e f o r e c a l l e d agent ive par e x c e l l e n c e . This
i s in essence a 'necessary and s u f f i c i e n t c o n d i t i o n s ' approach, which
if appl ie d to a s p e c i f i c area o f the vocabulary o f a language, such
as the one under i n v e s t i g a t i o n here, yields r a t h e r poor r e s u l t s . A
few verbs such as 'kinisao' (hunt) and 'kataftioko' (chase) can be
shown t o necessarily i n c or p or at e the feature of inten tionality and
therefore qualify as t r u l y a g e nt i ve . The r e s t would simply have to
be unmarked f o r t h i s f e a t u r e . Consi derati on o f a s i n g l e example w i l l
su ffic e at th is point:

- 145 -
(24) o j a m' s t a k s i b e v i j a t i j a r o [ V=t aksi bevo)
John i s t r a v e l l i n g t o Yaros.

It can be s a f e l y argued t h a t the sentence i s not ambiguous as to an

intentional and an u n i n t e n t i o n a l interpretation of ' taksibevo' , but
simply vague. So, ’ taksibevo' w ill be unmarked, or n eut ra l wi th
r espect to agentivity. The same would apply to 'y listrao' (slip),
though common sense would d i c t a t e a d i f f e r e n t i a t i o n between these two
at least: ' taksibevo' is usually understood as d e sc r ibi ng an
intentional act (although it a d m it t ed l y does not i nc or por at e an
' intentionality' f e a t u r e by n e c e s s i t y ) . The f a c t s o f 'y listrao' are
very d i f f e r e n t , however. The cases where i t i s not an acci dent to
slip are very r a r e . Ne ver the le ss , w i t h i n a system which works wi th
necessary and s u f f i c i e n t conditions, 'ylistrao' cannot be marked as
essentially different from ' taksibevo' in terms of agentivity; for
the former verb can also appear wi th a purposive c o n st r uc t io n :

(25) o kl oun y l i s t r i s e ke epese j a na kani t a pebj a na yel asun

the clown s l ippe d and f e l l to make the c h i l d r e n laugh.

Since (25) i s grammatical ' y l i s t r a o ' would have to appear in the same
category wi th ' piyeno' (go), 'fevyo' (leave) and 'taksibevo'
(travel), d e sp it e the fact that (25) reflects one of the least
c h a r a c t e r i s t i c uses o f the verb.

The problem w i t h all such accounts seems to be t h a t they restrict

a g e n t i v i t y and a gent ive verbs to the paradigm cases, i.e. those which
inv ol ve a prototypical understanding of the notion of agent onl y.
They consequently expect verbs t o be e i t h e r agent ive or non-agenti ve
on the basis o f whether the subj ec t they each accept is animate and
i n t e n t i o n a l or not . The r e l e v a n t f e a t u r es o f t h i s t r a d i t i o n a l notion
o f agent are more or l es s al re ady o f f e r e d in Lyons( 1 9 7 7 ) . The agent
i s understood as i n v o l v i n g an animate e n t i t y which i n t e n t i o n a l l y and
r espons ibl y uses i t s own for ce or energy t o bri ng about an event or
i n i t i a t e a process r e s u l t i n g in a change in the physical c o nd i ti o n or
location of some other entity. These properties indisputably
characterize what we may call a prototypical agent. There is no

- 146 -
indication, however, as to what happens in 'non-paradigm' cases,
where the r e l e v a n t f e a t u r e s are said to be separable from each o t he r .
We t h e r e f o r e need a more d e t a i l e d a na ly s is o f what t h i s p r o t o t y p i c a l
notion o f agent i nvolves and a t the same time a s p e c i f i c a t i o n of the
poi nts a t which d e v i a t i o n s from the protot ype can occur. This should
lead to a b e t t e r understanding o f verbs in d i f f e r e n t semantic domains
w it h r espect to t he pr op e rt y in quest ion. Such an understanding
i mpl ies raising the question of agentivity for each verb but not
expecti ng a yes-no answer.

Types of entities which can be regarded as closest to the

prototypical agent are n a t u r a l l y those perceived as having internal
energy, whose motion is not perceived as r e s u l t i n g from some ot he r ,
e x t e r na l cause and is therefore inferred to arise from within.
Natural e n t i t i e s (e.g. the sun or clouds) and n a t ur a l phenomena ( e . g .
rain, earthquakes, fire) are probably the c l e a r e s t cases, followed by
certain kinds of machines (e .g. v eh i c l es or computers).^ Objects
unable to act (move, function) on t h e i r own ( e . g . books, stones) and
t h e r e f o r e c h a r a c t e r i s t i c a l l y obj ect s not perceived as s el f- movi ng are
at the bottom of the scale. The priority of 'perceived' versus
'lo g ic al' cause cannot be too s t r o n g l y emphasized and examples from
the domain o f motion verbs are abundant in t h i s respect. Consider
all the motions a t t r i b u t e d to the unmoving sun: e . g . 'aneveni' (goes
up), 'vyeni' (comes o u t ) , 'y i r i z i ' (turns), 'fe w /' (leaves), 'p e fti'

(falls). Consider also the case of v eh i c l e s which seem to be

regarded as extensions o f human motion and are t h e r e f o r e compatible
w i th act ions ot herwi se attributed to humans and animals (including
'trexo' (run) i n MG). In short all these are e n t i t i e s wi th motion
per cei ved as self-generated; the most typical such entities are
humans, e s p e c i a l l y as they are also capable o f i n t e n t i o n a l a ct ion and
can be regarded as p r o t o t y p i c a l agents. Animals are obvi ousl y the
next step, since t hey also act i n t e n t i o n a l l y d e sp i te t he f a c t t h a t
sometimes their a ct i on s are attributed to instinct. Linguistic
evidence for this is the fact that they can appear as subjects o f
c le a rly intentional verbs such as ' k i n i y a o ' (hunt).

Control over one's motions is a characteristic closely linked to

intention (see Givon 1979) and animals are also understood as having

- 147 -
c ontr ol over t h e i r motions. After a l l , even human motions o f c e r t a i n
types such as 'tinazom ' (jump up) or 'tremo' ( t r em bl e ) are r a t h e r
automatic and not e a s i l y c h a r a c t e r i z a b l e as i n v ol v i ng e i t h e r control
or intent. In g e n er a l , every instance of controlled causation is
assumed t o i n vol ve intent. C l e a r l y the r e l a t i o n is not bilateral;
i nte ndi ng to do something does not n e c e s s a r i l y imply being also in a
p o s i t i o n to do i t . In t h i s context i t is t h e r e f o r e assumed t h a t both
'in ten t' and 'control' are only linked to animate (mai nl y human)
entities. If an a ct io n is characterized by 'c o n tro l', i.e. as the
result of an a ge nt ' s possessing control over the situation being
brought about, it i s assumed t h a t i t is also i n t e n t i o n a l (on the pa rt
o f the a g e n t ) . I t may, however, be the case t h a t although the agent
intends to bring an event about, s/he may not have f u l l ( or even any)
control over it. In the area under investigation this possibil i t y
w ill be discussed in connection wi th the d i s t i n c t i o n between d i r e c t
and i n d i r e c t c a u s at i v es .

It is therefore suggested t h a t f o r each verb i t can be i n v e s t i g a t e d

what sor ts o f agents it allows as subjects and the e x t e n t t o which
they are prototypical agents. Assigning a g en t - s t a t u s to entities
l ac k in g t he p r o t o t y p i c a l c h a r a c t e r i s t i c s o f agents may be considered
a process of metaphoric extension (Givon 1 98 4: 106 ). Since the
present investigation on the whole excludes the c ons id e ra ti o n of
metaphoric uses of MGMVs, i t i s important to s t a t e where the l i m i t is
drawn; f o r , c l e a r l y , cases such as:

(26) 7 skepsi tu petakse makria (V=petao)

his thought f le w f a r

although very interesting and e x p l a i na bl e within the framework of

Prototype theor y along the same l i n e s necessary f o r o t h e r types of
no n- p r o t o t y p i c a l uses, are not taken into account here. The sample
is restricted quite arb itrarily to p e r c e i va b l e physical motion and
concrete i n s t i g a t o r s of t h i s motion.

To r e c a p i t u l a t e what has a lr e ad y been said on the noti on o f agent,

the prototypical agent is understood as being human, clearly

- 148 -
perceived as the d i r e c t causer o f an event, i nt endi ng to br ing t h i s
event about and having absol ute c ontrol over i t s compl eti on. Wi thi n
the framework of Prototype t he o ry , agentivity is understood as a
cluster o f these p r o p e r t i e s r a t h e r than one d i s c r e t e feature. Each
pr oper ty is also understood as s c a l a r . Each subsequent p o i n t on the
scale marks a deviation from the p r ot o t yp e . The hierarchy within
each pr op e rt y does not seem to require further explanation. The
directionality of the scales constituting agentivity is considered
universal, but d i f f e r e n t languages are expected t o draw the d i v i d i n g
line at s l i g h t l y different points. The s p e c i f i c proposal presented
below o f how t h i s c l u s t e r o f scaled p r o p e r t i e s can be understood is
o f f e r e d in Givon ( 1 9 8 4 : 1 0 7 ) :

"a. Humanity: human > animate > inanimate > a b s t r a c t

b. Causation: d i r e c t cause > i n d i r e c t cause > non-cause
c. Volition: strong i n t e n t > weak i n t e n t > non -v o lu nt a ry
d. Cont r ol : c l e a r cont rol > weak c ontr ol > no c ontr ol
e. S al i ency : very obvious/salient > less obvious/salient >
unobvi ous/ nonsal ient"

Givon considers th a t ontogenetically (or at least ontologically) we

are de al i ng with one core p r op e rt y : obv iousness/ sal ience of cause.
His e xpl ana ti on i s worth quoting in f u l l :

"... a human i s c l o s e r to the ego, thus more f a m i l i a r and

obvious. D i r e c t causes tend to be p e r c e p t u a l l y more obvious,
occupying a c l e a r b o u n d a r y p o s i t i o n w i t h i n the chain (as also
does the e f f e c t , which is c a t e g o r i c a l l y coded as p a t i e n t ) .
I n t e rm ed i at e p oi n ts in the chain are l ess s a l i e n t . Strong i n t e n t
creates a h i ghe r p r o b a b i l i t y o f s u c c e s s , i . e . v i s i b l e e f f e c t .
D i t t o f o r strong c o n t r o l . " ( 19 84:107)

It is fa irly clear that this notion of agent is restricted in a

d i f f e r e n t d i r e c t i o n namely in subsuming a g e n t i v i t y under c a u s a t i v i t y .
For the moment, tra n s itiv ity and d i r e c t vs i n d i r e c t caus at ive s w i l l
be discussed in t h i s l i g h t and then an extension o f t h i s noti on w i l l
be attempted to cover i n t r a n s i t i v e s and MG mediopassives a l s o .

- 149 -
3.4 T r a n s i t i v i t y and a g e n t i v i t y

On the basis of this approach two prototypical c ondi ti ons are

p o s tu l at e d on which tran sitivity depends, i.e. the transitive
pr otot ype i n vol ve s:

( a ) the presence o f a v o l i t i o n a l , visib le, c o n t r o l l i n g cause

(i.e . a prototypical agent) as s ubj ec t;

(b) the presence o f a c l e a r l y v i s i b l e r e s u l t - r e g i s t e r i n g e f f e c t

or a p a t i e n t - o f - c h a n g e as o b j e c t .

Consequently prototypically transitive events within the field of

MGMVs w i l l be o f the type:

( 27) o makis si kose ena v i v l i o ( V=si kono)

Makis l i f t e d a book

(28) 7 mar i et a eval e t o m o l i v i st o t r a p e z i (V=vazo)

M a r i e t t a put the penci l on the t a b l e

(29) t a pebj a pet aksan ta x a r t j a mu st o patoma (V=petao)

the c h i l d r e n threw my papers on the f l o o r .

By c o n t r a s t w it h these sentences, sentence (30) w i t h a n o n - v o l i t i o n a l

s u b je ct and a non-typical o b j e ct (on the basis of c o n di t i o n (b)
above) w ill be characterized as registering an unpr ot ot yp ic al
transitive event, although it exhibits all t he syntactic
characteristics of t r a n s i t i v i t y :

( 30) o makis i be ena v i v l i o (V=v7epo)

Makis saw a book.

Following the practice established in the preceding c hapt er , the

c h a r a c t e r i s t i c s o f whole p r op os i t i on s are taken i n t o c o n si d er a ti o n in

- 150 -
order to reach a deci si on on the f e a tu re s of individual verbs
appearing i n them which are assumed to be the main c a r r i e r s o f these
characteristics. By j u xtaposi ng examples (27) and (30) whose subject
and o bj e c t NPs are i d e n t i c a l i t becomes t r a n s p a r e n t that w hi le 'to

vivlio ' ( t he book) in (30) i s not in the least affected by the

subject's a c t i on - and is t h e r e f o r e a very unpr ot ot yp ic al patient -
in (27) the o b j e c t ' s p o s i t i o n is d i f f e r e n t and may be regarded as a
prototypical i nstance o f a p a t i e n t .

The facts of t r a n s it iv it y I have sketched so f a r are in accordance

wi th GivoVs (1984) model. It should be obvious from what has
preceded t h a t , in his understanding, p r o t o t y p i c a l l y t r a n s i t i v e events
are col lapsed wi th prototypically agentive events. Thus a sentence
like (30) would be i n t e r p r e t e d as 1e s s - t h a n - p r o t o t y p i c a l l y t r a n s i t i v e
not only because its object NP is a n o n - p ro t o t yp i ca l patient, but
also because i t s subj ec t is a n o n -p r ot ot y pi c al agent since i t i s non-
v o litio n a l. This can be contrast ed with (27) where the opposite
c ondi ti ons hold. There can probably be l i t t l e o bj e c t i o n t o t h i s if
one considers the overwhelming importance o f the human a ct i o n model,
i.e. that a typical event is ling u istic a lly expressed in Indo-
European languages, at least, predominantly by reference to an
'actor/agent', t he a ct ion performed and the 'p atie n t/o b jec t/
recipient' of this a ct io n (see also Werner and Kaplan 1963, M iller
and Johnson-Laird 1976). What is perhaps more probl emati c is the
relation between prototypically agenti ve and prototypically
transitive sentences on the one hand and prototypically causati ve
ones on the o t h e r . This is not discussed in Givon, but on the basis
o f what has a lr e ad y been said one may assume t h a t a l l t h r e e c oi nci de .
Consider, however, the f o l l o w i n g sentences:

(27) o makis s ik o s e ena v i v l i o {V=sikono)

Makis picked up a book

( 30) o makis i 5e ena v i v l i o (V=v7epo)

Makis saw a book

- 151 -
(31) o makis k i t a k s e ena v i v l i o {V^kitazo)

Makis looked a t a book

Considering t r a n s i t i v i t y , c a u s a t i v i t y and a g e n t i v i t y s e p a r a t e l y , (27)

can be thought of as prototypically tran sitive, c aus a ti ve and
agent i ve ; (30) i s l ess p r o t o t y p i c a l l y t r a n s i t i v e than the o t he r two,
also less agent ive than either and non-causati ve; (31) is more
transitive and more a gent ive than (30) since it has a volitional
s ubj ect and less transitive than (27). However, precisely because
(31) contains a volitional subj ec t it is not equally non-agentive
wi th (30), although they are e q ua l l y n on - ca us at iv e. In shor t,
d i st an ce from the pr otot ype does not seem to me t o be a t parallel
poi nts along the t h r e e dimensions in question. In t h i s sense i t can
be thought t h a t t r a n s i t i v i t y c o n s t i t u t e s a c o n f l a t i o n o f the separate
dimensions o f a g e n t i v i t y and c a u s a t i v i t y .

In view o f what has preceded, it can be argued, however, that the

degree of causativity of transitive verbs depends mainly on the
degree o f d e v i a t i o n from the protot ype o f the p a t i e n t . Obviousness
of change in the o bj e c t can be shown to pl ay a more important r o l e
than a g e n t i v i t y o f the s ubj ect (for those cases a t l e a s t where the
two are not completely i n te r de p e n d e n t ) . N ot ic e , for i nst an ce , the
case o f MGMVs accepting a NP o f l o c a t i o n (see 3 . 2 . 1 , category I ) . I
have suggested that a p os si bl e measure/criterion for causativity
might be p a s s i v i z a b i l i t y . One might expect that the l ess affected
the ' p a t i e n t ' , the l ess possi bl e i t would be f o r i t t o appear as the
subj ec t o f a passive c o n s t r u c t i o n . Some o f the V-NP|_qc examples in
3.2.1, I can be c ontr ast ed w i t h V - P r e p - N P ^ c const ruct io ns ( wi th the
same verb) and seen as viewing the obj ect as more i mportant t o the
event than their V-Prep-NPLoc c ounter par ts in which the NPlqc is
viewed simply as a p o i n t o f r e f e re n ce f o r the s u b j e c t ' s l o c a t i o n or
movement (cf. Givon 1984:99). In view of these observations
r econs ide r now the f o l l o w i n g examples:

(32a) o jam's pibikse to f r a x t i { V=pi §ao)

John jumped the fence
vs

*(see examples (32) to ( 39)).

- 152 -
(32b) o j a m ’s p i b i k s e pano apo to f r a x t i
John jumped over the fence

(33a) t aksi bepse ( o l o ) ton kosmo ( V=t aksi bevo)

s/he t r a v e l l e d (all) the world
vs
(33b) t aksi bepse se ol o ton kosmo
s/he t r a v e l l e d t o / i n a l l the world
s/he t r a v e l l e d a l l around the world

There is c l e a r l y more than one way o f handling such cases. They may
be understood as parallel to a r ece nt development in MG syntax
whereby in very colloquial speech the preposition marking the
location of an object or the goal of the motion is completely
omi tted; e . g . :

(34a) p i y e n i Qesal oni ki [\l =pi yeno)

i nstead o f

(34b) p i y e n i s t i Qesal oni ki

s/he goes t o Sal oni ca.

Sim ilarly:

(35a) i ne p a r i s i (V=ime)

i nstead o f

(35b) i ne s t o p a r i s i
s/he i s in P a r i s .

This may be e xpl ai ned in terms o f the v e r b ' s having enough semantic
i nf or ma t io n - 'move towards g o a l ' in the case o f 'piyeno' (go) and
'p o sition in space' in the case of ' ime' (be) - to make the

- 153 -
p r e p o s i t i o n redundant. Notice, in t h i s r e sp e c t , t h a t in e i t h e r case,
i.e. directional or stationary, the preposition is the same (i.e .
'se') and t h a t e l i m i n a t i o n o f ' s t o ' / ' s t i ' (which is i n e f f e c t Prep. +
Article) a p p li es only a f t e r the two most general ( hi g h e s t taxonomic
level) verbs of the relevant categories, i.e. 'piyeno' and ' / me '
respectively.

An alternative view to redundancy in the case of 'pibao' and

' t a k s i b e v o ' in (32a) and ( 3 3 a ) , i s the one compatible with Prototype
t he o ry , namely const rui ng a locative as a patient object, i.e.
construi ng events as involving direct objects; but since these
obj ect s are not essentially affected by the subject's action, the
resulting events are very far from the transitive p r ot ot y pe . It
seems p l a u s i b l e t o assume t h a t p r e c i s e l y because the o bj e ct does not
become a prototypical enough patient in such cases, it cannot be
f u r t h e r promoted t o subj ec t p o s i t i o n in a passive c o n s t r u c t i o n , hence
(32a) does not have a corresponding:

(32c) * o fraxtis pibixtike (apo to j a n i ) (V=pi bj eme)

the fence was jumped (by John).

N e i t h e r does ( 3 3 a ) :

(33c) ( o l o s ) o kostnos t a k s i b e f t i k e (V^taksi bevome)

the (whole) world was t r a v e l l e d .

This change o f p e r sp e c t i v e may e x p l a i n why:

(36a) o j a n i s pr osper ase t o mayazi ( V=prospernao)

John passed by the shop

does not have a corresponding:

(36b) * t o mayazi p r o s p e r a s t i k e (apo to j a n i ) ( \ =pr osper nj eme)

the shop was passed by (by John)

- 154 -
w hi le (37a) does:

(37a) t o f i a t prosperase t i vol vo {\ =pr osper nj eme)

the F i a t overtook the Volvo

(37b) 7 vol vo p r o s p e r a s t i k e apo to f i a t [V^prospernj eme)

the Volvo was overtaken by the F i a t .

It can be said t h a t the car in s ubj ect p os i t i on in (37b) i s in t h i s

case understood as ' a f f e c t e d ' by having been overtaken by another car
(as, f o r i n st an ce , in a car r a c e ) .

A c l e a r e r case i s perhaps presented by (38) vs ( 3 9 ) :

(38a) o j a n i s b i e s x i s e t o basos (V=5 7 a s x i z o )

John c ro ss e d/ tr av e rs ed the wood

(38b) to basos b i a s x i s t i k e (apo to j a n i ) ( V=bi asxi zome)

the wood was c ro ss e d/ tr av e rs ed (by John)

(39a) ena pot ami b i a s x i z i to basos ( V=5 7’ a s x i z o )

a r i v e r t ra v e r s e s the wood
vs
(39b) t o basos b i a s x i z e t e apo ena potami (V=b7 asxizome)
the wood i s t ra v er s ed by a r i v e r

The examples in (39) register something permanent and important in

connection w i t h the wood; i t is i n a way cut in two by the r i v e r . It
is therefore plausible to consider t h a t in this latter case it is
closer to the prototypical patient as it is more affected by the
s ub j e c t . Consequently (39a) is closer to the transitive prototype
than (38a) and t h i s could be the reason why i t has a corresponding
p a s sive.^ The c r u c i a l thing to n ot i ce here is that it is not the
properties of the s u b je ct , but rather the condi ti on of the obj ect
which i s d e c i s i v e here; w h i le the subject in (38a) is both animate
and v o l i t i o n a l , the s ubj ect in (39a) i s n e i t h e r .

- 155 -
For s i m i l a r reasons (40a) has a p a r a l l e l in (40b) and (41a) i n (4 1 b) :

(40a) o exBros p e r i k i k l o s e t i n p o l i (V=perikiklono)

the enemy surrounded the c i t y

(40b) 7 p o l i p e r i k i k l o Q i k e apo ton exQro ( V=per i ki kl onome)

the c i t y was surrounded by the enemy

(41a) ena s i n e f o kapnu p e r i k i k l o s e t o s p i t i { V= p e r i k i k l o n o )

a cloud o f smoke surrounded the house

(41b) t o s p i t i p e r i k i k l o Q i k e apo ena s i n e f o ( V=per i ki kl onome)

kapnu
the house was surrounded by a cloud
o f smoke.

A p os si bl e e xp l an a ti o n of the f a c t s of ' §iasxizo' and ' p e r i k i k l o n o '

i s t h a t John in (38b) is a single e n t i t y , occupying a p oi n t in space
a t any given ti me, and t h e r e f o r e unable to s t r e t c h a l l the way across
a wood (t her eby dividing it into two p a r t s ) at a given moment in
time. On the ot he r hand, since a r i v e r extends c o nsi de ra bl y f u r t h e r
on a h o r i z o n t a l plane than 'John', it can e a s i l y d i v i d e a wood i n t o
two (as in (39)). Li kewi se, 'enemy' in (40) designates a whole
collection of ind ividua ls, who can t h e r e f o r e e a s i l y form themselves
into a c i r c l e , i.e. the e n t i t y has e x t e n t . The e s s e n t i a l f a c t s about
'wood' in (39), 'c ity ' in (40) and 'house' in (41) (by comparison
with 'wood' in (38a)) is the relationship in which they stand to
something wi th e x t e n t , a r i v e r and the enemy/the smoke r e s p e c t i v e l y .

In any case, the qualities o f the subj ect are not decisive in the
difference between (38b) and ( 4 0 b) . The object seems t o be more
seriously affected in (40a, 41a) and therefore closer to the
tran sitive pr otot ype than (38a). For the same reason it is also
closer to the c au sa ti ve p r ot ot y pe . Most importantly, animacy and
intentionality of the s ubj ect are not crucial in any o f t he cases
just discussed. Appar entl y the degree o f causativity rests mainly

- 156 -
wi th the degree of 'p a tie n tiv ity ' of the affected object.
'prospernao' (overta ke), 'perikiklono' (encircle, surround),
'biasxizo' (traverse) + NP|_qc> although they are not causati ves of
motion, are nearer the c au sa ti ve prototype ( i n the sense of change of
c o n di ti o n of the object) than ot he r [NPj_oc] verbs which cannot
passivize. This d i f f e r e n c e depends wholly on how much o f a p a t i e n t
the o bj e c t i s and t h e r e f o r e on q u a l i t i e s o f the s ubj ect ot he r than
animacy and i n t e n t i o n a l i t y .

A number o f poi nts r a i s e d i n e a r l i e r sections ( 3 . 1 , 3 . 2 , 3 . 3 ) w i l l be

taken up in what f ol l ow s and examined in the light of this
'prototy pe' understanding of tra n sitiv ity and agentivity. It is
considered t h a t w i t h i n t h i s framework more l i g h t can be c ast on the
r e l e v a n t issues, i.e. t r a n s i t i v e s which are not caus ati ve s o f motion,
the r e l a t i o n between p h o n o l o g i c a l l y i d e n t i c a l c a u s a t i v e / n o n - c a u s a t i v e
p a i r s , d i r e c t vs i n d i r e c t c a u s a t i v i t y , agentivity.

3.5 D i r e c t vs i n d i r e c t causati ves

The distinction between direct and indirect c ausati ves has been
a l re a dy mentioned. It w ill be argued here that it constitutes
perhaps the best way to investigate the relationship between
c a u s a t i v i t y and a g e n t i v i t y . Consider f i r s t the verb ' x o r e v o ' (dance)
which can be used in a v a r i e t y o f ways and occupy d i f f e r e n t poi nts on
a scale of t r a n s i t i v i t y and c a u s a t i v i t y . It w ill be looked a t in
some d e t a i l as i t c o n s t i t u t e s a good example o f the i n s u f f i c i e n c y o f
f or ma l , syntactic characteristics and restricted notions of
agentivity.

(42) 7 marieta x o re v i to xoro t i s k i l j a s {\=xorevo)

M a r i e t t a dances ( t h e ) b e l l y dance

(43a) o filipo s xorevi ti marieta

P h i l i p dances M a r i e t t a
P h i l i p 'dances w i t h Marietta.

- 157 -
(43b) o f i l i p o s x o r e v i me t i marieta

P h i l i p dances wi th M a r i e t t a

(44) 7 dada x o r e v i t o mo r o s t a yonata tis

( t h e ) nanny dances the baby on her knees

The d i f f e r e n c e s between the var ious poi nts on the scal es in question
can be captured by r e f e r e n c e t o the notions a lr e ad y discussed. As
the verb i s b a s i c a l l y an i n t r a n s i t i v e of motion, none o f the NPshere
construed as d i r e c t obj ect s are p r o t o t y p i c a l patients. In (42) the
o bj e c t NP, f a r from being a p a t i e n t , is in e f f e c t s i m i l a r in fu nc t io n
t o a manner ( a d v e r b i a l ) specification. In (43a) it could be thought
of as der ived from a PP whose NP i s in that case construed as a
direct object; the o bj e c t is affected more than in (43b) which
descri bes the same e ve nt , in a way, in that the in itiativ e of
performing the event is attributed to the s u bj e c t ; intention is
probably shared but viewed as mainly stemming from the s u b j ec t ; weak
cont rol only can be a t t r i b u t e d t o the s ubj ect as the o b j e c t - p a t i e n t
must necessarily also use its sel f- movi ng mechanism ( i n the
a p p r op r i at e manner). In t h i s sense the causer i s n o n - s a l i e n t . The
main p o i n t here is that the event i s described as an i nstance of
i n d i r e c t c au sa ti on. In (43b) the o b j e ct appears w i t h i n the PP and is
clearly viewed as l es s affected than in (43a); all the f a c t or s
respons ibl e for the execution of the act/event (i.e. intention,
control, cause, responsibility, in itia tiv e ) are viewed as shared by
subj ec t and object. The question of causation does not arise.
Finally, in (44) the o bj e c t i s a p r o t o t y p i c a l patient; i t need not be
using i t s sel f- movi ng mechanism in the l e a s t , and although i n t e n t i o n
may be common t o s u bj ec t and o b j e c t , it i s c l e a r l y a t t r i b u t e d t o the
former, who is also the d i r e c t causer o f the event and has c l e a r and
obvious c ontr ol over it. This is an i nstance of a very salient
causer, a prototypical agent: she is both pe rcei ved and wholly
r e spons ibl e f o r t he eve nt. By c o n t r a s t wi th ( 4 3 a ) , t h i s is a case of
d i r e c t c au sati on.

This particular verb also cla rifie s the s ta t us of the causatives

belonging to cat egory I (in 3.2 ). It seems quite reasonable to

- 158 -
regard cases o f what I c a l l 'phono!ogically i d e n t i c a l ' causative/non­
caus at ive p a i rs as a type o f extension o f meaning. The r e s u l t of a
process o f e l i m i n a t i n g p re p os i t i o n s i s a d i r e c t o b j e c t in place of a
PP (constr ui ng a l o c a t i v e as a p a t i e n t ) . Assuming t h a t causati ves of
motion of the type in question are der ived from the corresponding
intransitives, one can think of them as the result of a process
parallel to the one just mentioned in connection with PPs:
' c o n s tr u in g an i n t r a n s i t i v e as a t r a n s i t i v e ' . The close l i n k between
t he various uses of ' xor evo' just discussed can be seen as an
e x t ens ive example of such a process. The process starts wi th a
genuine i n t r a n s i t i v e as in ' o f i l i p o s x o r e v i st o b a l e t o t i s operas'
( P h i l i p dances a t the b a l l e t o f the Opera). There f o l l o w cases such
as the examples provided (42) - (44) where the l a s t one, at least,
i nvolves a direct caus at ive of motion. There is, of course, no
question of r egardi ng these various instances of ' xorevo' as
'homophones'. S im ilarly for ' /c7 7a 0 ^' (ro ll), 'str/vo^' (turn),
'vuljazo2 ' (sink) and the remaining verbs of category I of MG
causati ves o f motion.

In discussing the facts of direct vs indirect causati ves it seems

useful to compare them w it h exp licit c a us a t i ve s , expressed in MG
through verbs such as ' kano' (make), ' anagazo' (force) and the l i k e ,
and l i n k the issue wi th the a g e n t i v i t y sca les .

Consider the d i f f e r e n c e s between e x p l i c i t causati ves ( h enc ef or th EC),

and lexical ones (he nc ef or th LC), comprising direct and indirect
causati ves (he nc ef or th DC and IC r e s p e c t i v e l y ) :

EC (45a) 7 v r o x i mas ekane na yi r i s u me p i s o {V=yirizo)

the r a i n made us come/go back

LC (45b) ??7 v r o x i mas y i r i s e p i s o

the r a i n turned us back

- 159 -
EC (46a) 7 proelasi t u exQru ekane t o strato na (V=ip oxoro)
ipoxorisi

the advance of t he enemy made the army

retreat

LC (46b) *7 proelasi t u e x Qr u i p o x o r i s e to strato

the advance of the enemy ' r e t r e a t e d ' the army

EC (47a) *ekana ta vivlia na a n e v u n s t o rafi (\l=aneveno)

I made the books go-up on the s h e l f

LC (47b) anevasa t a vivlia sto rafi ( V=anevazo)

I t oo k - u p / p u t - u p the books on the s h e l f

EC (48a) o j a n i s mas e k a n e na anevume p a n o x o r i s na (V=aneveno)

t o 0 e 7 7*

John made us go-up u p s t a i r s w ithout

wishing to

LC (48b) ?o j a n i s mas a n e v a s e p a n o me t i s fones tu (V=anevazo)

xoris na t o Qeli

John t ook-up us u p s t a i r s wi th his screams wi thout wishing to

John made us go u p s t a i r s wi th his screams wi thout wishing to

The main poi nts to no ti c e here can be summed up as f ol l o w s :

( a) Explicit causati ves may have causers that are f a r from

prototypical agents, e.g. (45a) and (46a) which are also
explainable as instances of ' re as on' rather than 'cause',
although t he d i f f e r e n c e is hard to analyse i n these terms. Even
when the causer is human, as in (48a), s/he may l a c k all the
remaining c h a r a c t e r i s t i c s o f a p r o t o t y p i c a l agent. I n t e n t i o n may
be absent and c ontr ol very weak. Responsibility is a subjective
c h a r a c t e r i s t i c i n t h a t i t may be a t t r i b u t e d by the speaker-causee
'u n ilaterally' to the person s/he understands as the causer.

- 160 -
 E x p l i c i t c ausati ves are t y p i c a l l y used when the causee i s not a
prototypical patient (e .g. (47a) is completely b l oc k e d ) , since
they r e q u i r e t h a t s/he must also take action, i.e. be a c t i v e l y
i nvolved in the event and therefore at least share the
r e s p o n s i b i l i t y f o r i t s completion.

(b) The d i s t i n c t i o n between DCs and ICs is not as c l e a r - c u t as may be

thought, except if one concentrates on p r o t o t y p i c a l i ns tan ce s.
Prototypically DCs should exhibit the characteristics of
prototypical transitives, e.g . (47b) which is a clear case of
'm anipulative' causati on: the causer has a l l the properties of
the p r o t o t y p i c a l agent and the causee is a p r o t o t y p i c a l patient.

(c) ICs are much more n a t u r a l l y paraphrasable wi th a corresponding EC

expression than DCs a r e, e.g. ( 4 8 a ) - (48b) vs (47a)-(47b). The
reason i s obvious: parallel t o ECs t h e i r prototypical instances
i n v ol ve the causee's active participation in t he event; the
causer is attributed responsibility for the action; s/he must
also exhibit at least 'weak intent' (e.g. (48b)) but e xe r ci se
only weak c o n t r o l . The causer i s , however, p r o t o t y p i c a l l y human;
hence the u n a c c e p t a b i l i t y of (45b) and ( 4 6 b ) .

Individual MG causati ves of motion w ill d iffer as to the

possibilities they a l l o w and may be seen as ranging from cases very
close to DC to cases o f metaphoric use.^

In view of what is s t at e d under (b) above, verbs which can be

p r o t o t y p i c a l l y used as DCs w i l l have a high degree o f a g e n t i v i t y . In
the f i e l d under i n v e s t i g a t i o n , t h i s involves verbs such as:

' plisiazc>2> (approach), 'sfr/vo^' (turn), fy ir iz o 2' (turn),

'p e r if e r o ' ( t ak e around), 'bizo' (stick), ' p e t a o^ ' (t hrow) and others
(see L i s t I V ) .

Examples:

- 161 -
(49a) 7 n i n i p l i s i a s e t i n k a r e k l a st o t r a p e z i (\ = p l i s i a z o )
Ni ni approached the c h a i r to the t a b l e

(49b) *7 n i n i p l i s i a s e ton j a n i st o t r a p e z i
Ni ni approached John to the t a b l e

(49c) ??<? aeras p l i s i a s e ta f i l a s t i n porta

the wind approached the leaves t o the door

The subject-causer should have clear control, be a direct and

absolutely salient causer besides being animate and v o l i t i o n a l and
usually human, hence the unaccepta bility of (49c). By i m p l i c a t i o n
the p a t i e n t has t o be also a p r o t o t y p i c a l one, in the sense o f having
none of these properties; hence the unacceptability of (49b).
Together w i t h 'metakomizo2f (transport f u r n i tu r e , etc ., w h i le moving
house), 'ylistrao2 f (slip) and 'kikloforo' (circulate), 'plisiazo'
(approach) features at the top of the agentivity scale for MG
causati ves o f motion, as can be a t t e s t e d in L i s t I V. A l l these verbs
are shown t o accept only animate causers (and sometimes only human),
scoring very high for each of the properties jo in tly comprising
agentivity. The only pr ope rt y not a l l o t t e d a speci al column in L i s t
IV is ' o b v i ou s ne ss / sa l i e nc e of c aus e' . It is considered t ha t
salience is in fact a characterisation of the 'sum t o t a l ' of the
remaining p r o p e r t i e s .

A number o f o the r verbs such as ' i psono ' ( r a i s e ) , ' x a mi l o n o ' ( l o w e r ) ,

' par amer i zo' (push aside), ' me t a f e r o ' (transport), 'bizo' (stick),
'petao^' (throw), 'tinazo' (shake u p / o f f ) , 'rixno' (t hrow) are shown
to de v i a t e s lightly from the prototype of agentivity by a ll owi ng
e n t i t i e s o th e r than human and animate as s u bj e c t s , e . g . ;

(50) o aeras xami l ose t i s ombreles ( V=xami l ono) •

the wind lowered the umbrellas

(51) 7 x j o n o s t i v a b a t us e r i k s e kat o ( V=r i xno)

the snowball threw them down.

- 162 -
We are s t i l l , however, in the area o f high a g e n t i v i t y since in the
case o f animate subj ects these verbs are marked f o r implying simple
intention, clear c ontr ol and d i r e c t cau satio n .^ The few possi bl e
exceptions are marked in L i s t IV with an x r a t h e r than a + f o r the
relevant p r op e r t y , as an indication that such uses are marginal ^,
e .g .:

(52) ?o a s t i n o m o s p a r a m e r i s e t u s f i t i t e s me (\l=paramerizo)
tis fones tu

the policeman pushed the students aside wi th

hi s screams.

Thus 'paramerizo' (push asi de) w ill be r e l u c t a n t l y marked (x) for

'weak' besides 'clear' control, to take care also of uses such as
e xe m p l i f i e d in ( 5 2 ) .

Verbs marked as accepting subjects e x e r c i s i n g either strong or weak

control as well as a ll ow ing the possibility of IC are 'ferno'

(bring), ' p /v en o^ ' ( t ak e to), ' apomakrino' ( t ak e away), ' sikono'

(raise, lift), 'vazo' (put), ' vyazo' ( t ak e o u t ) , ' anevazo' ( t a k e up),
e tc ., e .g .:

(53a) p/ye t a r u x a s t o k a Q a r i s t i r i o {\l=piyeno)

s/he took the clo the s t o the laundry

is p r o to ty p ic a lly agentive, while:

(53b) p i ye t i 07 a t u s t o 0 e a t r o , opos t o n ixe (\l=piyeno)

parakalesi

he took h i s aunt t o the t h e a t r e , as she had

asked him to

i s less so, because although the agent i s high on the 'animacy' scale
and has simple i n t e n t i o n , he has 'weak c o n t r o l ' o f the s i t u a t i o n : the
responsibility for the act is shared w it h the causee and the
in itia tiv e i s a t t r i b u t e d to the l a t t e r . Thus t h e r e i s good reason to

- 163 -
presume t h a t the causee is actively participating i n the event (at
least using her sel f- movi ng mechanism). Not ice also some instances
of 'f e rn o ' (bring):

(54a) t a t r ena f er nun t r a v ma t i e s {V^f erno)

the t r a i n s br ing wounded (persons)

(54b) 7 Balasa e f e r e s k u p i b j a s t i n p a r a l i a
t he sea brought l i t t e r to the beach

w i t h s ubj ects low on the 'animacy' sca le; and

(54c) e f e r a t o f i l o mu st o p a r t i
I brought my f r i e n d to the p ar ty

which is sim ilar to (53b) in that the o bj e c t is again animate,

v o litio n al, using h i s / h e r sel f- movi ng mechanism, e t c . , in short q u i t e
an u np ro t o t y p i c a l patient. These l ess agent ive verbs pose an
i n t e r e s t i n g problem. Should they be understood as i n v o l v i n g i n d i r e c t
causati on in uses such as those e x e m pl i fi ed in (53b) and (54c) above,
or not?

The d i v i d i n g l i n e between d i r e c t and i n d i r e c t causation has g e n e r a l l y

been assumed to be presence vs absence of 'manipulation' (cf.
S hi bat ani 1976a, b, where direct causation is called ' ma n i p u l a t i v e
causation') and the borders are assumed to be clear. It seems,
however, t h a t t h e r e i s g ra d a t i o n w i t h i n both the m a ni p ul at iv e and the
non- ma ni pul at i ve kind and in fact a line l ea d i n g from explicit
c ausati ves to p r o t o t y p i c a l l y d i r e c t ones. Cruci al features in t h i s
connection seem t o be: 'c o ntro l', 'accompaniment' and ' i n i t i a t i v e ' ,
all l e ad i ng to a specification of the degree of responsibility of
causer and causee, and degree o f s al i ence o f cause.

In the absence o f a supporting c onte xt , it i s d i f f i c u l t t o prove the

v a l i d i t y o f these obs er v at io ns . An attempt w i l l be made, however, a t
a careful a n a l y s i s o f r e l e v a n t sentences in the hope o f showing t h a t

- 164 -
what are here posi ted as 'crucial fe atures' are indeed rightly so
characterized and e s p e c i a l l y t h a t none o f them, taken i n d i v i d u a l l y ,
is s u f f i c i e n t . Consider the f o l l o w i n g examples:

(55a) mas ekane na pame st o p a r t i me t i n a p i l i ( V=pi yenoi )

o t i a l j o s ben 0a mas k s a n a m i l i s i
s/he made us go t o the p a r t y t h r e a t e n i n g
t h a t otherwi se s/he would not t a l k to us agai n.

This is an i nstance o f an EC i n vo lv i ng intransitive 'piyenoj' (go) .

The causer does not participate physically in the executi on o f the
actual caused eve nt . S/he does not accompany the causees. There i s
good reason to b e l i e v e , though, t h a t s/he has ' s tr on g in te n t'. The
remaining examples i nv ol ve transitive-causative uses of 'piyeno2f
( t ak e t o ) .

(55b) mas p i y e st o p a r t i me t i n a p i l i o t i ( V^p/yeno^ )

a l j o s ben 0a mas k s a n a m i l i s i
s/he took us to the p a r t y t h r e a t e n i n g t h a t
otherwise s/he would not t a l k t o us again

In (55b) the causer is attributed strong intent; s/he has the

in itiativ e in the change-of-location event and also accompanies the
causees who can be only a t t r i b u t e d weak i n t e n t and no i n i t i a t i v e .

(55c) p i y e t i mi t e r a t u st o j a t r o opos ton ixe (V=piyeno 2 )

parakalesi
he took his mother to the doctor as she had
asked him t o do

The main difference here seems to be that although the causer

accompanies the causee, both ( s tr ong) intent and in itia tiv e are
a t t r i b u t e d to the l a t t e r .

(55d) piye t i yata st o j a t r o ( V- pi yeno 2 )

s/he took the c at to the doctor

- 165 -
Here, o f course, the causer possesses all the characteristics of a
prototypical agent and the causee is more or l ess a prototypical
patient. In essence (55d) is p a r a l l e l to (53a) 'piye ta ruxa sto

kaQ aristirio' ( s / he took the c lothes to the l a u n d r y ) .

Examples (55a- d) are meant as an i l l u s t r a t i o n o f the gradedness of

the phenomenon o f c a u s a t i v i t y , which can be d i r e c t l y linked to the
facts of a g e n tiv ity (as a lr ea dy pointed out at the beginning o f t h i s
chapter). A d irect/indirect dichotomy does not seem very
illum inating. The trad itio n ally used feature of 'accompaniment',
although useful f o r d i f f e r e n t purposes, is not very important in t h i s
p a r t i c u l a r context and can be r eplaced by the pr ope rt y o f ' c o n t r o l ' .
No t ic e , f o r i ns t an ce , that (55a) is an i nstance o f very weak control
and low s al i e n c e on the p a r t o f the causer, that (b) and (c) seem to
d i f f e r s i g n i f i c a n t l y in terms o f control ( the former one i n v o lv i n g of
weaker c ontr ol than the l a t t e r ) , whi le (d) i s c h a r a c t e r i z e d by strong
control. The suggestion here i s t h a t ' c o n t r o l ' seems to take care of
'accompaniment' in a more r e v e a l i n g way, as i t can l i n k d i r e c t l y with
s a l i e n c e o f cause. It i s a p p l i c a b l e under the f o l l o w i n g c ondi ti ons:
the act can only be performed by the a ct or being p h y s i c a l l y present
and a ct i n g until its completion; at the same t im e, no active
p a r t i c i p a t i o n is r equi re d on the p a r t o f the moving o b j e c t ( p a t i e n t ) .
These c ondi ti ons are c h a r a c t e r i s t i c o f s i t u a t i o n s considered here to
be p r o t o t y p i c a l l y under the c l e a r control o f the causer. If, or to
the extent that, they are satisfied, the verb is considered more
agent i ve (than othe rwi se ) in this r e sp e ct . Thus, f o r instance,
'ipsono' (raise) w ill get a hi gher score than ' sikono' (raise),
'anevazo' (p ut/take up), 'ferno' (bring), etc., all ot he r things
being e qual , since cases parallel to (55a), (55b) and (55c) are
impossible in the case o f ' ipsono' (raise); but not in the case of
the latte r verbs. This means t h a t it can only f u nc t i on as a c l e a r
case of a direct c a u s a t i v e . The s a li e nc e of acauser as a
prototypical agent cannot be disputed in cases such as that
e x e m p l i f i e d by a l l i nstances o f ' i p s o n o ' ( r a i s e ) .

Situations us u a l l y expressed w i t h an EC ( e x p l i c i t causative), where

the causer may be simply i ssuing an o r de r, making a suggestion,
asking the causee t o perform an a c t , g i v i n g d i r e c t i o n s , and the l i k e ,

- 165 -
are in cer ta in cases of MGMVs also expressible through a LC (le x ic a l
ca u sat ive ).

Examples of such causative verbs are: ' sikono1 (raise), ' kaQi zo'
(make s.o. sit), ' v az o ' (p ut ), ' vyazo' (take out), ' k s apl ono' (make
s.o, l i e down), ' anevazo' (take up/make s.o. go up), ' kat ev az o ' (take
down/make s.o. go down), ' pernao 2 ' ( p a s s e s ) ’ 'strivo2 ' (tu rn ^ s),
etc. Char act er isti c examples of the d i f f e r e n t possibi1i t i e s are:

(56) 7 baskala sikose ton j a n i ston pinaka (M=sikono)

the te ac he r ' stood up' John t o the blackboard
the t e ac h e r made John stand up in f r o n t of t h e blackboard

(57) 7 b as ka la e v y a l e t o n j a n i ekso ( V = w az o )
the t e ache r ' t o o k out' John
the t e a ch e r made John go out/asked John to go out

(58) 7 n a z i e s t i s a n tu s kratumenus ston t i x o (V=stino)

the Nazis stood t he pr isoner s at the wall

(59) evale t a p e b j a mesa j a na m i n k r i o s u n (V=vazo)

s/he put the c h i l d r e n i ns ide so t h a t they would
not catch cold
s/he made them go i n s i d e so t h a t they would not catch cold

(60) perase tu s ksenus s t o saloni (V=pernao)

s/he passed the guests t o the lounge
s/he showed the guests i n t o the lounge

(61) zitisan apo t o n o b i y o na t u s k a t e v a s i (V-kate vazo)

stin proteleftea stasi

they asked from the d r i v e r to take them down

at one stop before the l a s t
they asked the d r i v e r t o put them o f f
a t one stop before the l a s t .

- 167 -
In L i s t IV such verbs are marked f o r a ll owi ng both strong and weak
c ontrol and d i r e c t as wel l as i n d i r e c t cause, j u s t l i k e those of the
previous cat egory, i.e. ( take t o ) , ' ferno ' (bring). They
have to be s ingl ed out , however, as a ll owi ng a f u r t h e r d e v i a t i o n from
the protot ype o f d i r e c t c a u s a t i v i t y , in t h a t the causer need not be
obvi ousl y e x e r c i s i n g c ontr ol over the event, f o r the reasons o f f e r e d
in the discussion of the examples ( 5 5a - d ) . It would perhaps be
desirable to have more poi nts on the 'control' scal e to mark t h i s
deviation, since ' c o n t r o l ' , in my understanding o f the term, i nvolves
more than one condition (as pointed out in the di scussion of
g r a da t io n and c o n t r o l ) .

The most i n t e r e s t i n g case in t h i s connection is perhaps presented by

those MGMVs which are even c l o s e r to the EC p r o t ot y p e, having the
speci al characteristic that they are most naturally r e p l a c ea b l e by
e q u i v a l e n t EC c ons t ru ct io n s ( u n l i k e the immediately preceding group).
Significantly, these are r a t h e r c l e a r cases o f extension o f meaning,
e.g.:

(62a) mas t r e x i s t a mayazja ( V=t rexo 2 )

s/he makes us run to t he shops

(62b) mas t r e x i j a f o t o t i p i e s
s/he makes us run f o r photocopies

(63) mas s e r n i st on ipobromo [V=serno)

s/he drags us to the horse-races

(64) me e s t i s e m i s i ora ekso apo t o sinema (V=stino)

s/he ' s t o o d ' me f o r h a l f an hour outsi de the cinema
s/he kept me w a i t i n g outsi de the cinema f o r h a l f - a n - h o u r .

Not ice that (62a) and (62b) may be e i t h e r + / “ accompaniment, w h i le

(63) i mpl ies accompaniment, and (64) i mpl ies the complete absence o f
the causer from the location of the caused event; (64) is most

- 168 -
commonly used f o r an appointment made ( p r i o r t o the caused event) and
not kept by the s u b j e c t .

S al i ence o f cause is very low in these examples because cont rol is

very weak. If the causer i s simply i ssui ng an o r de r , s/he is l ess
s a l i e n t than a d i r e c t causer, e . g . (62)-(64) asopposed to ' c a r r y i n g '
someone/something; if s/he i s not even present during the execution
o f the a c t , s/he is even less s a l i e n t , e.g. (6 2a-b), (64) as opposed
to (6 3 ). The verbs are not understood l i t e r a l l y ; although they imply
'change-of-location' ( ' se r n o ' (drag), ' trexo' (run)), or 'change-of-
p o s i t i o n ' o f the causee { ' s t i n o ' (make s . t h . stand, p l a c e ) ) , probably
only the f e a t u r e o f 'speed' is r e t a i n e d in ' t r e x o ' (run), and ' s e r n o '
(drag) r e t a i n s l i t t l e more than ' u n w i l l i n g n e s s to perform t he a c t ' or
'a c tin g as i f dragging one's f e e t a l o n g ' . As f o r 'stino' (make s . o .
stand), it probably keeps the feature of 'having an u p ri g ht
position', although even that is not absolutely necessary; the
predominant characteristic in such uses i s 'bei ng a t a fixed po i nt
u n w i l l i n g l y or f o r a l onge r period than one intended t o ' . It i s also
interesting to n o t i c e that although the causer i s low on the scales
o f contr ol and cause, and perhaps n e u t r a l / u n s p e c i f i e d f o r degree of
intent, s/he i s very high on the animacy sca le ( u s u a l l y human), e . g . :

(65) *7 v r o x i mas e s i r e s p i t i mas ( \i=serno)

the r a i n dragged us home

(66) *o bi nat os i l j o s mas e s t i s e s t i n p a r a l i a ( V = st /no)

the strong sun ' s t o o d ' us on the beach

(67) * t o a f t o k i n i t o mas t r e x i s i n e x i a s t o s e r v i s ( V= tr exo^ )

the c ar makes us run conti nuousl y t o the ' s e r v i c e .

This shows the importance o f understanding a g e n t i v i t y as a c l u s t e r of

p r o p e r t i e s ; d e v i a t i o n from the pr otot ype o f a g e n t i v i t y i s expected to
t ak e place in any one (or more than one) of its individual
properties. I t also gives an idea o f how extensi on o f meaning may be
understood to work, i.e. as d e v i a t i o n from t he p r o t o t y p e .

- 169 -
It has been a l re ady s ta t ed here t h a t only physical motion i s taken
i n t o c o n si d e ra t i o n and no f i g u r a t i v e uses of MGMVs. Yet cases l i k e :

(68) o Borivos apomakrine tus p erastiku s (\l=apomakrino)

the noise drove away the passers-by

which are taken into account since they i n vol ve physical change-of-
location of the object and a concrete causer-subject can well be
thought o f as a step in the process o f meaning e xt ensi on. Consider
(69) as a paraphrase o f ( 6 8 ) :

(69) 7 perastiki a p o m a k r i n B i k a n apo t o Borivo/ (V-apotnakrinome)

eks e t i a s tu Borivu

passers-by p ul l e d / w e n t away because o f the noi se.

It seems p l a u s i b l e t h a t the NPs expressing cause or reason w i t h i n a

PP { [ a p o t o 0or7'vo]PP, [eks e tia s tu Borivu]??) in (69) are construed
as agents and become subjects in sentences l i k e (68) where the causee
i s consequently 'demoted' from the s ubj ect p o s i t i o n i t holds in (69)
to d i r e c t o bj e c t p o s i t i o n in ( 6 8 ) , Givon's model does not r a i s e t h i s
point, but I consi der i t q u i t e i n t e g r a t a b l e in any form o f Prototype
t h e or y . It is q u i t e conceivable t h a t every step on a scale marking
deviation from the protot ype of a transitive, c au s a t i v e , agentive
c o n st r uc t io n should be understood as a step towards extension of
meaning and may br ing about a rearrangement o f the r o l e s / f u n c t i o n s o f
the d i f f e r e n t arguments o f a p r e d i c a t e , focusing some and defocusing
o t h er s . The resulting const ruct io ns w ill in effect present the
'same' event viewed from a d i f f e r e n t angle.

3.6 A g e n t i v i t y measurements f o r i n t r a n s i t i v e s

In what has preceded, some arguments were presented a gai nst a

complete c o n f l a t i o n of causativity and a g e n t i v i t y and in favour of
consi deri ng only transitives as candidates for membership in the
category o f c au s a t i v e s , a t l e a s t f o r methodological reasons. I t has
also been suggested t h a t the degree o f a g e n t i v i t y o f causati ves of

- 170 -
motion could be 'measured' on the basis o f Givon's (1984) a g e n t i v i t y
s ca les . List IV shows how the verbs so examined compare w i t h each
ot he r w it h r espect t o t h i s p a r t i c u l a r dimension. The only pr oper ty
(among the cluster of properties jo intly constituting agentivity)
which is not attributed a separate column is salience. It is
considered t h a t e s p e c i a l l y in the absence o f any s p e c i f i c context ual
i n fo r m a t i o n , the relative s al i e n c e of the 'causer' is in fact
reflected in the o v e r a l l p i c t u r e o f how a gent ive the verb i s on the
basis o f a c o n si d er a ti o n o f the remaining p r o p e r t i e s .

A further deviation from Givon's a nal ys is is im plicit in what has

al re ady been said concerning a g e n t i v i t y , namely t h a t i n t r a n s i t i v e s of
motion also exhibit different degrees of agentivity. It is
considered therefore that one-argument predications can be also
c at egor i ze d as more or l ess a gent ive on the basis o f p r o p e r t i e s such
as animacy, in tention and c o n t r o l . The r e s p o n s i b i l i t y of the a ct or
is now measured in relation to his/her own a c t i v i t y , which is not
transferred to any ot he r entity. The pr o pe r ty of causation is
inapplicable in t h i s case, especially as i t i s presented in Givon's
s ca l es , i.e. in terms o f d i r e c t and i n d i r e c t cause, so t h e r e is no
column corresponding t o i t in L i s t V which i ncludes non-causatives of
motion. Transitives d esc ri bi ng the motion o f t h e i r logical subject
are considered along wi th intransitives o f motion. The d i s t i n c t i o n
drawn is in fact between causati ves of motion (which can only be
transitives) and non-causatives (which can be e i t h e r t r a n s i t i v e , e . g .
' k i n i y a o ' (run a f t e r , hunt) and ' k a t a b i o k o ' (c hase), or i n t r a n s i t i v e ,
e.g. ' t r e x o j ' ( run) and ' aneveno ' (a s ce n d) ) . The assumption i m p l i c i t
here is that causati ves are to be compared for the purposes of
agentivity measurements wi th ot he r causati ves onl y; and,
consequently, non-causati ves with o t he r non- ca us at iv es. Consider,
firs t, sample sentences wi th ' k a t a b i o k o ' (chase):

(70a) 7 ast i nomi a k a t a b i o k i t on b r a p e t i ( V=kat abi oko)

the p o l i c e are chasing t he f u g i t i v e

(70b) t o p e r i p o l i k o k a t a b i o k i t on b r a p e t i
the p a t r o l c ar i s chasing the f u g i t i v e

- 171 -
(70c) ??7 sferes katabiokun ton bra pe ti
the b u l l e t s are chasing the f u g i t i v e

(70d) *7 petres katabiokun ton brapeti

the stones are chasing the f u g i t i v e .

C l e a r l y the verb scores high on the 'animacy' s ca le; the same a ppl ies
to the 'intent' and 'control' s ca les . It har dl y needs to be
demonstrated w it h examples that to chase somebody one must have a
strong i n t e n t i o n t o do so and c l e a r control o f the s i t u a t i o n ; it is
most u n l i k e l y t h a t the event i s in accordance wi th t he wishes o f the
chased person or t h a t the l a t t e r has any a c t i v e involvement in the
accomplishment o f the a c t i o n .

I t can be shown t h a t ' t r e x o ' (r un) is a step lower in a g e n t i v i t y than

verbs like 'katabioko' (chase) and 'kiniyao' (hunt, run after); on
the 'animacy' s ca l e, besides animates, v e h i c l e s , and n a tu r al f or c es ,
it also accepts certain kinds of non-self-moving subj ec ts such as
liquids and mass obj ect s moving under the effect of gravity or
impetus, e . g . :

(71) to nero trexi mesa s t o n k u v a ( V = tr e xo )

the water i s running i n t o the bucket.

There are l ess grounds for postulating 'strong intent' for 'trexo'

(run) than f o r 'katabioko' (chase) and ' k i n i y a o ' (hunt , run a f t e r ) .

It is, however, concei vabl e t h a t activities requiring a relatively
g r e a t e r amount o f energy t o accomplish than others should be marked
for marginally exhibiting ' s tr o n g intent' besides ' s im pl e intent'
(hence 'trexo' r ece ive s x in the relevant box). The basis for
marking t h i s way verbs such as ' trexoj' (run), ' skarfalono' (cl imb)
and 'ko liba o' (swim) is the fact that they yield quite unnatural
sentences i f combined wi th expressions such as ' x o r i s loyo' ( f o r no
reason), 'x o ris na kseri ja ti' ( wi th ou t knowing why), etc. It is
f e l t t h a t t he re i s a d i f f e r e n c e a t t h i s poi nt from ' p e r p a t a o ' (walk),
f o r i n st an ce , which i s q u i t e compatible wi th such expressions and is
therefore only marked for 's impl e in te n t'. Notice a lso that

- 172 -
7p e r p a t a o ' i s hi ghe r than ' t r e x o 7 on the animacy scal e as i t r equi re s
animate subj ects onl y, but 'vabizo' (s t ep ) and ' porevome' (go on
foot), although very s i m i l a r to ' p e r p a t a o 7 in most r e s p e c t s , are even
hi gher on the animacy scale as they r e q u i r e only human s u b j ec t s. All
these verbs are generally a gent ive but not for the same reasons;
hence they appear at different levels of agentivity. No t ic e , for
i ns tan ce , that ' ormao 7 ( bu rs t in/on) has to be marked for ' s tr ong
intent' r a t h e r than 's impl e i n t e n t ' but is lower on the animacy scale
since v e h i c l e s and n a tu r a l f or ces (besides animates) can also appear
as its subjects. Most of these verbs in the more or l ess
p r o t o t y p i c a l l y a gent i ve group are marked f o r 'clear control'. It is
fe lt, however, t h a t ' y o n a t i z o ' (kneel), 'kaBome' (sit) and ' ksap l on o'
(lie down) should also be marked for marginally a ll o wi n g 'weak
c o n t r o l ' since t hey can appear i n sentences such as:

(72) yonat i se apo t o k s a f n i k o x t i p i m a ( V=yonat i zo)

s/he kneeled under/because o f the sudden blow
or
(73) z a l i s t i k e k s a f n i k a ke ksapl ose st o patoma ( V=ksapl ono)
s/he suddenly f e l t d i z z y and l a y on the f l o o r

where ' epese st a y o n a t a ' (s/he f e l l on h i s / h e r knees) and ' epese st o

patoma ' ( s/ he fell (fla t) on t he floor) are implied, respectively.
It also seems reasonable to mark 'periplanjeme' (roam around) for
'n o n -in te n t' as w el l as 's impl e i n t e n t ' to take care o f uses such as:

(74a) p e r i p ] a n i B i k a n apo laBos p l i r o f o r i e s { V^ per i pl anj eme)

they roamed around because o f wrong d i r e c t i o n s

(74b) p e r i p i a n i B i k a n s t i n p o l i j a na pe r a s i i ora
they roamed around ( i n ) the town t o pass the t i m e6 .

So the reason why ' periplanjeme' s till figures within t he agenti ve

group is its high score for animacy. Other verbs comparable in
agentivity to t he ones discussed are: 'busulizo' (crawl (as of a
baby)), ' xorevo' (dance), 'xoropibao' (hop), ' serjanizo, sulatsaro'
(walk a r o u n d / a b o u t ) .

- 173 -
The c l a s s i f i c a t i o n o f verbs i n t o groups i s meant only as a r e f l e c t i o n
of what their 'characteristic image' is. There are clearly no
borders between these 'groups'. So ' t r e x o ' (r un) and ' p i b a o ' (jump)
require the active physi cal involvement o f the. moving s u b j e c t , yet
are lower than the ot he r verbs j u s t discussed in terms o f animacy.
Although they may be f e l t as p r o t o t y p i c a l l y a g en t i v e , they are shown
t o be l ess so. These two verbs are in a way the l i n k with the next
group, c o n si s t i n g of verbs generally understood as u ns pe ci fi ed for
i n t e n t i o n and t h e r e f o r e ' vague' as to a g e n t i v i t y . I am r e f e r r i n g to
MGMVs such as:

'ta k s ib e v o ' (travel), 'e p is tre fo j' (return), 'p lis ia z o i' (approach),
'b ia s x iz o ' (traverse), ' aneveno' (ascend), 'p iy e n o j' ( g o) , 'erxom e'
(come), 'fe v y o ' (leave), ' beno' (enter), 'y iriz o j' (turn)

and others which a ll o w f o r s i t u a t i o n s where the moving e n t i t y may be

lower on the 'animacy' scale than 'human' and may execute the
described motion while being carried, for i ns t an ce , i.e. not
necessarily using its s el f- movi ng mechanism. To take care of this
l a t t e r p o s s i b i l i t y these verbs are marked f o r both ' s t r o n g ' and 'weak
co ntrol'. It is interesting to notice, however, t he differences
which they e x h i b i t .

A small number among them: 'e p is tre fo j' (return), 'y iriz o 3 p is o '

(come back), 'p lis ia z o ' (approach), 'ta k s ib e v o ' (travel), 'b ia s x iz o '

(traverse) are compatible wi th q u a l i f y i n g phrases of the kind ' x o r i s

na to k a ta la v i' ( w i th ou t knowing/realizing it) and can thus be
s i ngl ed out as m a r g i n a l l y a ll o wi n g 'non-intent' and 'no control'.
Unl i ke ' t a k s i b e v o ' (travel), f o r i nst an ce , 'p iy e n o j' (go) and ' f e v y o '
(leave) yield very odd sentences in such environments. What is
perhaps more i n t e r e s t i n g , i f the moving e n t i t y is understood as being
unconscious, or in any way completely l a ck i ng contr ol or intention
w i t h r espect to t he motion descr ibed, 'erxom e' (come) sentences such
as (75) are f i n e , but ' p i y e n o ' or ' f e v y o ' ones, such as (76) and (77)
are not:

- 174 -
(75) s tis 5 irQan i travm a ties s to nosokomio {V=erxome)

at 5 o ' c l o c k the wounded came to the hos pi ta l

(76) lls tis 5 e fiya n i travm a ties apo t o nosokomio (V=fevyo)

a t 5 o ' c l o c k the wounded l e f t the hos pi ta l

(77) lls tis 5 p iya n i travm a ties s to nosokomio {V^piyeno)

a t 5 o ' c l o c k the wounded went to the h o s p i t a l .

In case ' i tra vm a tie s' ( t he wounded) are completely l a c k i n g control

or intention, sentences wi th 'tu s p ira n ' ( t h ey took them away) and
'tu s p iv a n ' ( t hey took them (to)) w ill re pl ac e (76) and (77)
respectively. The equivalent c au sa ti ve c o n st r u ct i o n 'tu s eferan'

(they brought them) need not be substituted for (75). For this
reason i t seems p l a u s i b l e to mark ' e r x o m e ' (come) as also accepting
'no c o n t r o l ' and 'non-intent' (besides 'clear control' and 'simple
in t e n t '),

Notice the d i f f e r e n c e between ' p / y e n o j ' (go) and ' f e v y o ' (leave):

(78) o li 7 k r a t u m e n i e f / y a n apo t i s fH a ke s a verof; (V=feyyo)

tus p ira n s tis 6

all the pr isoner s l e f t the prison ' A v e r o f f ' ;

they took them (away) a t 6

(79) o li 7 kratum eni p iya n s tis fila k e s averof; (V=piyeno)

tus e fe ra n /p iy a n s tis 6

a l l the p r i so ner s went t o the prison ' A v e r o f f ' ;

they brought them/took them t h e r e a t 6.

I t seems t h a t ' p i y e n o ' r e qu i r e s a t l e a s t 'weak c o n t r o l ' w h i l e ' f e v y o '

has to be marked f o r m a r g i n a l l y also accepting subj ects w i t hout any
control over the event.

A possi bl e e xp l ana ti on for these facts could be t h a t 'p/yeno' (go)

focuses on the whole j ourney rather than its beginning or its

- 175 -
end( 1 ike 'come' and ' l e a v e ' ) . I f the speaker focuses on onl y the end
point, 'con tro l' i s perhaps l es s r e l e v a n t than when s/he focuses on
the whole o f i t .

Predictably, a compl etel y d i f f e r e n t o v e r a l l p i c t u r e i s given by verbs

such as 'k a tra k ila o ' (roll down) or 'tre m o ' ( t r em b le ) which f i g u r e
q u i t e low on a g e n e r a l i z e d a g e n t i v i t y s c a l e , f o r a number o f reasons.
For i nstance ' k a t r a k i l a o 1 ( r o l l down) scores very low on 'animacy' as
it is b a s i c a l l y r e s t r i c t e d to non-sel f-movi ng obj ect s (i.e . entities
w it hout a s el f- movi ng mechanism or not using i t t o perform the motion
in q u e s t i o n ) . It i s also a t the bottom o f both the ' i n t e n t ' and the
'co ntrol' sca les , as it i s completely i ncompatible w i t h anynotion of
intention, e.g.:

(80) * k a tra k ilis e ja na p j a s i ti bala [V = k a tra k ila o )

s/he r o l l e d down t o catch the bal l

as well as wi th any noti on o f the s u b j e c t ' s having c ont rol over the
event. The d i f f e r e n c e s within this group are very slight. Three
verbs: 'y lis tr a o j' (slip, slid e ), 'v iB iz o m e 1 (sink) and 'p e fto '

(fall) can be m a r g i n a l l y marked f o r ' weak/simple intent' and 'weak

control' (besides 'non-intent' and 'no c o n t r o l ' , which c h a r a c t e r i z e
the p r o t o t y p i c a l uses o f these v e r b s ) , e.g.:

(81) Iv iB is tik e sto 5r o s e r o n e ro {V=viBizom e)

s/he sank i n t o the cool water

i s acceptabl e f o r a number o f n a t i v e speakers.

3.7 Modern Greek mediopassives and passives as i n t r a n s i t i v e s

In the l a s t category discussed in the previous s ecti on a number of

verbs ending in '-ome' (hence '-o me ' verbs) are i ncluded: 'v iB iz o m e '

(sink), ' gremizome' (fall down/to p i e c e s ) , ' sorjazom e' (fall flat on
the ground). MG ' - o w e ' verbs are mediopassive in form and, wi th the

- 176 -
obvious exception of 'deponent verbs' such as 'e rxom e' (come) or
' apomakrinome’ (move away), are trad itio n ally juxtaposed to active
form '-o ' verbs as implying that their subj ec t is affected rather
than a f f e c t i n g some ot he r e n t i t y .

I f t h i s view were accepted a t i t s f a c e - v a l u e , '-ome' verbs should not

have been considered f o r a g e n t i v i t y measurements a t a l l ; they could
be simply s ingled out as d istinctly non-a ge nti ve. It has seemed,
however, on closer inspection, that the facts of these verbs
concerning agentivity are not as simple as that, nor is their
behaviour uniform in t h i s respect. It i s suggested here t h a t these
facts should be analysed and expl ai ned along s i m i l a r l i n e s to those
a ppl ied to o the r one-argument predicates, i.e. 'ordinary'
intransitives. In order t o show the n e ce s si ty o f c onsi der ing ' - om e'
verbs together with intransitives for agentivity measurements, it
seems i mportant to look b r i e f l y a t the r e l a t i o n s between passives and
ot he r related constructions. Such relations do not seem to be a
s t r a i g h t f o r w a r d m a t te r even i n languages which have been much more
thoroughly analysed than MG, e.g. Engl ish. It seems best to s t a r t
the discussion with a brief c ons id e ra t io n of English passives in
r e l a t i o n to c a u s a t i v i t y , t r a n s i t i v i t y and a g e n t i v i t y .

A sentence such as 'the k n i f e was broken' is offered by M i l l e r and

Johnson-Laird ( 1 9 7 6 : 5 1 8 - 9 ) as an example o f a di minished passi ve; the
missing a ct or is understood, 'br oken' is an adjective and the
omission i s s y n t a c t i c . I t seems p l a u s i b l e to argue t h a t t h i s absence
o f an e x t er n al causer o f the event i s the l i n k between passives and
intransitives. Passive sentences have been recognized as
intransitive ones in t he literature for independent reasons (e.g.
Perlmutter 1 983). But the schema I envisage concerning causative-
transitives, intransitives and passives i nvolves different steps
along a scal e o f 'di mi ni she d importance o f an e xt e r n a l causer' and is
c l o s e r t o the approach adopted by Shibatani (1985) who c h a r a c t e r i z e s
p a s s i v i z a t i o n as a process o f ' a gent -de foc usi ng' ( p . 830). The actual
scal e I have in mind may be e x e m p l i f i e d as f o l l o w s :

- 177 -
 (a) (82) 'John broke the g l a s s ' (Transitive-causative)
Causer: present and s p e c i f i e d

(b) (83) 'Someone broke the g l a s s ' ( Tr ansitive-cau sative)

Causer: present but uns pe ci fi ed

(c) (84) 'The glass was broken' (di mi ni shed Passive)

Causer: only i mpl ied

(d) (85) 'The glass broke' (Intransitive)

Causer: compl etel y absent

This understanding o f passives l i n k s also w it h the f a c t t h a t they do

not generally express agents overtly, i.e. agentless or diminished
passives are f a r more numerousthan those w i t h o v er t agents even in
those languages which do al low o v er t expression of an agent. The
implications are p a r t i c u l a r l y i mportant for present purposes, since
cases o f passive sentences accepting an [apo NP] p r e p o s i t i o n a l phrase
e q u i v a l e n t to by-phrases f o r English are q u i t e r a r e in c o l l o q u i a l MG
and a l l the '-o/ne' verbs under i n v e s t i g a t i o n are complete in terms o f
valency w it h only one NP. The po i nt I am making concerning the
mediopassive form i s t h a t wi th MGMVs, a t l e a s t , even in the case of
genuine passi ves, the agent is not so much s y n t a c t i c a l l y omitted as
s em a n t i c a l l y downgraded and p r a g m a t i c a l l y brought out o f focus in the
sense o f S hi bat ani (1985:837). If MG passives are so interpreted,
the relation between genuine passives and potentials can be also
brought out . MG p o t e n t i a l s are expressed with mediopassive verb-
forms, e . g . :

(8 6 ) ben b j a v a z e t e afto to v iv lio (V=bjavazome)

i t - i s - n o t - r e a d t h i s book
t h i s book i s unreadable

(87) b en t r o y e t e afto to fa i {V=troyome)

i t - i s - n o t - e a t e n t h i s food
t h i s food i s i n e d i b l e ^ .

- 178 -
Far from r e g i s t e r i n g something about an e x t er na l agent, sentences
(8 6 ) and (87) seem to me to focus on some pr ope rt y of their
respective s u bj ec t NPs. What is perhaps more illuminating,
'spontaneous e ve nt s' are expressed in MG both through i n t r a n s i t i v e s ,
e.g .:

(8 8 ) afto to vazo s p a z i e fkola { M= s p a z o)

t h i s vase breaks e a s i l y

- which seems to me like a clear statement about a proper ty

attributed to the subj ec t NP - and, in the absence of such an
i n t r a n s i t i v e , through a mediopassive-form verb, e.g.:

(89) afto to ilik o ka ta stre fe te e fkola (y ~ k a t a s t r e f o m e )

th is material i s destroyed e a s i l y .

There would seem to be no real d i f f e r e n c e between (89) and p o t e n t i a l s

as e xe m p l i f i ed in ( 8 6 ) and ( 8 7 ) .

A number o f examples w i t h MGMVs o f mediopassive form w i l l h op e f u l l y

constitute c o rr o bo r a t i n g evidence for the c lai m that passives are
sim ilar to intransitives in the sense that the involvement of an
e xt e rn a l agent i s not invoked:

(90) p etres tinazonde ston a e r a apo t i n k o rifi ( V=tinazome)

tu ife s tiu

stones a re- thr own-up i n t o the a i r from the top

o f the volcano

(91) b ix tik e ena a g a Q i s t o p o b i mu (\l=bizome)

i t - w a s - s t u c k a thorn i n t o my f o ot
a thorn got i n t o my f o o t

(9 2) 7 staQmi tu neru ipsoQ ike ( M=ipsonome)

the l e v e l o f the water w as - r a i s e d / r o s e

- 179 -
(93) o p ira v lo s sikoQ ike (V = s i k o n o m e )
the r oc ke t was-r ai sed (rose)

(94) siko Q ik e sko ni ( V=sikonome)

dust rose

(95) to s p iti g r e m i s t i k e me t a x r o n j a {V^gremizome)

the house was-pulled/brought-down w i t h time

(95) to a fto k in ito xodike s ti la s p i mexri ta tzamja ( V=xonome)

the c ar was e n gu l fe d / s t u c k / w e n t i n t o the mud
up t o the windows

(97) ta nera tis lim n is ta ra xtika n (V^tarazome)

the waters o f the l ake wer e- di st ur be d

(98) to pe§i ine ane vasmen o s t o Sendro { \ l ~ a ne v az o me)

the c h i l d i s mounted/gone up on the t r e e

the c h i l d is on the t r e e (as a r e s u l t o f having gone up
the t r e e ) .

(99) 7 m arieta ine skarfalom eni sto fra x ti (V=skarfalonome)

M a r i e t t a i s ' c l i mb e d' on the fence
M a r i e t t a has climbed on the fence (and is
now on the fence as a r e s u l t ) .

It seems compl etel y counterintuitive to consi der any of t he above

cases anal ysabl e as: fx moves as the r e s u l t of the a ctivity o f an
e xt e r n a l causer'. Despit e the f a c t t h a t most o f the subj ec t NPs in
these examples are i nanimate and some of them are also non-self-
moving, no e x t er n a l agent is viewed as r esponsi bl e for the motion
either. They are not r e l a t a b l e t o ' cor respondi ng' a c t i v e sentences,
e .g .:

(90') kapjos tin a z i p et res apo t i n k o rifi tu ife s tiu

someone i s throwing up stones from the top of the volcano

- 180 -
(91') kapjos ebikse ena a ga Q i sto pobi m
someone stuck a thorn i n t o my f o o t .

In (95) and (96) the ( e x t e r n a l ) cause of the motion i s i ne f f e c t

gravity. But the language chooses to express what is actually
observable. G r a v i t y is not a s a l i e n t enough cause o f motion. So no
agent i s r e g i s t e r e d in the l e a s t ; no agent i s even i mpl ied in these
cases. Not ice t h a t in (92) an a c t i v e form i n t r a n s i t i v e can re pl ac e
the verb, i.e. ' aneveno' (go up). S i m i l a r l y in (91) ' b i k e 1 (went i n )

can r eplace the mediopassive 'b ix tik e ', Sentences (92) to (94) are
s i m i l a r t o cases l i k e :

(100) o iljo s v y i k e p i s o apo t o v uno (V=vyeno)

the sun came out from behind the mountain

where ' v y e n o ' (go/come out) can be also replaced by a number o f other
intransitives, e.g. 'p e ra s e ' (passed) and ' a ne vike ' (went up). As
has been a l re ady poi nted out ' gremizome' (be pulled down) can be
replaced by ' e p e s e ' (fell down) in (95) and ' x o Q i k e ' (was engul fed)
by ' v i Q i s t i k e ' (sank) in ( 9 6 ) .

Examples (90) to (97) are i nstances of such dramatic 'agent-de-

focusi ng' t h a t they move to category (d) (example ( 8 5 ) ) o f the scale
sketched at the beginning of 3.7 where absolute absence of the
e xt er nal causer is i m p li ed . Examples (98) and (99) re pr es ent extreme
cases o f t h i s situation; (98) cannot be r e l a t e d to ' a n e v a s t i k e ' (was
taken up); it simply impl ies 'to pebi anevike sto bendro' (the child
went up onto the tree) and views the resulting situation as more
permanent than a simple event, i.e. as having some o f the p r o p e r t i e s
of a state (see Chapter 2 ) . As f o r (99), the mediopassive form o f
' ska rfa lon o' (climb), namely ' skarfalonotne' (be cli mbe d) , does not
occur in any form ot he r than the Participle ' skarfalom enos' except
marginally in potentials, e.g. 'ben ska rfa lon ete afto s o vraxos'

( t h i s rock is ' u n c i i m a b l e ' ) .

The inevitable conclusion must be that the borders between

intransitives and passives are at least fuzzy. This i m p li es that

- 181 -
some p r o t o t y p i c a l or 'core' instances of passives can be d e te ct ed.
What has been called here a 'genuine passi ve' refers to cases
f u l f i l l i n g the f o l l o w i n g c r i t e r i a : (a) mediopassive form o f the verb;
(b) moving NP a f f e c t e d , i.e. theme = p a t i e n t ; (c) e x p l i c i t l y present
second argument: a NP w i t h i n an [ apo NPjPP expressing the entity
responsi bl e f o r the motion o f the t he me - p a t i e n t ; (d) the NP of the
[apo NPJPP possessing the c h a r a c t e r i s t i c s o f the p r o t o t y p i c a l agent.
One could then f u r t h e r assume t h a t the verbs occur ri ng in genuinely
passive const r uct io ns a r e , almost by d e f i n i t i o n , at the bottom o f the
agentivity scale. Hence i d e n t i f y i n g the l e a s t agent ive MGMVs could
amount to d i s cov er ing which verbs can occur in genui nel y passive
sentences. Yet even t h a t would be an u n r e a l i s t i c task; the marked
tendency of MG t o r e pl a c e passive const ruct ions w it h e q u i va l en t
a c t i v e ones r e s u l t s in the number o f actual 'genuine passi ves' being
extremely small w i t h i n any s p e c i f i c semantic f i e l d (e.g. MVs). Most
a c t i v e verbs have a corresponding passive form which i s a u t o m a t i c a l l y
formed by analogy to common-use passives (e.g. ' m etaferome' (be
transported)), but it is virtu ally impossible to find natural
environments for them, i.e. acceptabl e sentences c on ta i ni ng these
verb forms. What is even more intriguing, [apo NPJPPs can also
express 's ource', e.g. ' apom skrinB ike apo tus astinom us' (he moved
away from the policemen) w i l l not be i n t e r p r e t e d as a passive a t a l l ,
although the verb is mediopassive in form ( ' apomakrinome' (move
away)) and ' i astinosni' (t he policemen) has a l l the c h a r a c t e r i s t i c s
of a prototypical agent.

We are therefore left wi th one-argument predications where the

mediopassive ending simply predisposes for a non-agentive
understanding o f the verb8 . It can be shown, however, that '-o/ne'
(mediopassive form) verbs e x h i b i t d i f f e r e n t degrees o f a g e n t i v i t y and
that even ge nui nel y passive verbs can be more agenti ve than some
a c t i v e form i n t r a n s i t i v e s .

A c l o s e r l ook a t s p e c i f i c ' -ome’ verbs is intended t o support these

points. Consider f i r s t ' gremizome' (be p u l l e d down, fall down) and
' v i B i z o m e 1 (be sunk, s i n k ) , which have a lr eady been used as examples.

- 182 -
(101a) to s p iti g re m istike apo t i n astinom ia (V^gremizome)

the house was p ul l e d down by the p o l i c e

( 1 0 1 b) to s p iti g re m istike apo t u s sismus

the house was brought down by the earthquake

( 1 0 1 c) to s p iti g re m istike apo mono t u e p i b i ita n eripomeno

the house f e l l by i t s e l f because i t was d e c r e p i t

(102a) to p l i o v iB is tik e apo t o n k a p e t a n j o (V = viB izom e)

the boat was sunk by the capt ai n

(102b) to p l i o v iB is tik e apo t i n trik im ia

the boat sank by/because o f the tempest

(102c) to p l i o v iB is tik e apo m j a v la v i s ti m ixa n i tu

the boat sank because o f some damage i n i t s engine

The c l o s e r the cause to the p r o t o t y p i c a l agent the more l i k e l y it is

t h a t we are d e a l i n g wi th a genuine passi ve. Hence the ( a) sentences
are the only ones where, as a lr e ad y pointed out, 'g re m izo m e ' and
'viB iz o m e ' cannot be r eplaced by ' p e f t o ' (fa ll) and ' v u l j a z o ' (sink)
respectively. In the absence o f any [ apo NP] s p e c i f i c a t i o n t h e r e i s
no way of knowing whether the verbs in question are passive or
intransitive. The degree o f a g e n t i v i t y o f both verbs is q u i t e low,
s i m i l a r t o t h a t o f t he s em a n t i c a l l y r e l a t e d a c t i v e form i n t r a n s i t i v e s
just mentioned and for precisely the same reasons; the motions
described are very sim ilar and the e n t i t i e s undergoing them are o f
the same type. What is particularly interesting here is that
' viB izom e' may m a r g i n a l l y occur in an i n t e n t i o n a l environment, u n l i k e
’ v u lja z o ', and i s therefore slightly hi gher than the latter on the
s c a l e.

The c l e a r e s t case o f a ge nui nel y passive verb, which i s hi ghe r on the

agentivity scal e than a number of active form intransitives, is
'obiyum e' (be led). I consider both (1 03) and (104) e q ua l l y

- 183 -
prototypical i nstances of this verb and equally ge nui nel y passive
sentences.

(103) 7 exm a lo ti o biyiQ ikan apo t u s f r u r u s se [V-obiyume)

stratopebo sige nd rose os alisobem eni ke
pano se f o r t i y a

the hostages were l ed by the guards to

a c o n ce n tr a ti o n camp, in chains and
on l o r r i e s

[subject: human, n o n- v ol u n ta r y, no c o n t r o l / v e r y weak c o n t r o l , non­

cause]

(104) o ipu ryo s e k s o te rik o n t i s liv iis {V=obiyume)

o b iyiB ike apo t o n ib jo ton p ro B ip u ry o

s tin eBu s a s i n e b r i a s e o n o p o s i x e z i t i s i
the m i n i s t e r o f f o r e i g n a f f a i r s o f Libya
was l e d by the Prime M i n i s t e r h i m se l f
to t he conference room j u s t as he had asked ( t o be)

[subject: human, simple i n t e n t , weak c o n t r o l , i n d i r e c t cause]

The p o i n t , o f course, is t h a t ' obiyum e' is q u i t e compatible wi th the

notion o f the s u b j e c t ' s i nte ndi ng the event t o take p l a ce , having at
l e a s t some cont rol over i t and some r e s p o n s i b i l i t y f o r i t , as wel l as
having the i n i t i a t i v e f o r the accomplishment o f the a ct ion de scr ibed.
It is q u i t e e v id e nt t h a t i t should t h e r e f o r e appear somewhere in the
middle o f the a g e n t i v i t y s ca l e, e s p e c i a l l y as i t n e c e s s a r i l y r equi re s
animate (and u s u a l l y human) s ubj ec ts . Needless to say, i n t r a n s i t i v e s
such as ' k a t r a k i l a o ' (roll down), ' v u lja z o ' (sink) and the l i k e are
much lower on a g e n e r a l i z e d a g e n t i v i t y s ca le.

A compl etel y d i f f e r e n t p i c t u r e i s given by ' a p o m a k r i n o m ' (move away)

which should be considered a clear case of a 'deponent verb', a
characterization tra d itio n ally given t o verbs which were thought to
be passive in form but active in meaning. It must be poi nted out

- 184 -
here, t h a t the r e l u c t a n c e t o adopt the term w h o l e - h e a r t e d l y does not
imply t h a t this is an e n t i r e l y wrong way o f t h i n k i n g of them. The
problem i s , rather, that precisely because degree o f a g e n t i v i t y and
therefore 'how a c t i v e a verb is' is not a given but a m a t te r for
investigation, it i s not in the l e a s t c l e a r from a semantic poi nt of
view a t which p o i n t we l eave deponent verbs and move i n t o t he area of
t r u e passives.

The verb ' apomakrinome' is compatible wi th both simple intent and

non-intent; also wi th c l e a r and weak c o n t r o l , f o r the same reasons as
active form i n t r a n s i t i v e s like 'e p is tre fo i' (return) or 'p lis ia z o j'

( approach) . Thus (105) and (106) are e q u a l l y prototypical uses of

' apomakrinome':

(105) apom akrinBike y r iy o r a apo t i fo tja (V=apomakrinome)

s/he moved away q u i c k l y from the f i r e

(106) x o ris na t o kseri apom akrinotan sin e x o s

apo t o m e r o s p u iB e le na p a i

w i t h ou t knowing i t , he was conti nuous ly moving away

from the place he wanted to go

It is also c l e a r t h a t one can 'move away' not only u n w i l l i n g l y but

also as a passive passenger on a conveyance. For these reasons
' apomakrinome' ends up rather high on the agentivity scal e of
intransitives, scoring s i m i l a r l y to ' e r x o m e ' (come).

A number of examples have al re ady been given of ' -ome' verbs

(sentences (90) t o ( 9 7 ) ) which f un c t i o n as i n t r a n s i t i v e s in the sense
that no external causer can possi bl y be invoked, although the
subj ects in the s p e c i f i c examples are inanimate and o f t en n o n - s e l f -
moving. I t was claimed t h e r e t h a t no corresponding a c t i v e sentences
exist (al though t h e r e do e x i s t completely u n r e l at e d a c t i v e sentences
with the corresponding active verb f or ms). This i m pl ie s that in
these cases we cannot talk of diminished/truncated passive
constructions.

- 185 -
The verbs in question ar e: 'ip s o n o m e ' (be risen, rise), ’ sikonome'

(be rise n /lifte d , rise), 'x onom e' (become immersed/stuck/engulfed

in), 'b iz o m e ' (become stuck i n ) , 'tin a z o m e ' (be thrown up, jump up),
'p e tayo m e ' (be thrown up, l eap up, jump). A typical example of how
t h i s l a s t verb is understood here i s :

(107) petayonde f l o y e s apo t a p a r a B i r a {V=petayome)

a r e- be ing- t hr ow n flames out o f the windows

flames jump out o f the windows.

Needless t o say, t he passive and i n t r a n s i t i v e English verbs proposed

as translations are not meant as an indication that the verbs are
ambiguous in MG, The p oi n t here i s that they can be shown to
function sim ilarly to active form intransitives like 'x a m ilo n o j'

(lowerj^jjp), for i n st a n c e , and r e ce i v e sim ilar agentivity scores to

them. A r e l e v a n t example w i t h ' x a m i l o n o ' is (108):

(108) ta f i l a tu fitu xam ilosan apo t i n z e s t i

the leaves of the p l a n t d r op pe d / ' l ow er e d' from the heat .

It is interesting to notice th a t although the cause ( o r perhaps the

reason) is e xp lic itly given in (108) and t he re does exist a
corresponding c au s at iv e 'x a m ilo n o ' (lowerjp) e.g. 'xa m ilo s e ta m atja

tu ' (he lowered hi s e y e s) , no corresponding active-causative

c o ns t ruc t io n e x i s t s for (108), i.e .:

(109) *7 z e s t i xa m ilose ta fila tu f i t u

the heat lowered the leaves o f the p l a n t

is ungrammatical. Appa re nt ly 'the heat' in (108) is not deemed a

s a l i e n t enough cause to acqui re the s tatus o f an agent.

The l a s t examples o f mediopassive verbs i n vol ve ' a n e v a z o m e ' (be taken

up) and 'katevazom e' (be taken down). The complete paradigm is as
follows:

- 186 -
(a) t r a n s i t i v e (c ausative): ' anevazo' ( t ak e up), ' katevazo' ( t ake
down)

(b) i n t r a n s i t i v e a c t i v e form: 'an eve no ' (go up), 'kateveno' (go down)

(c) ( i n t r a n s i t i v e ) mediopassive form: ' anevazome' , ' k a t e v a z o m e ' .

Although (c) verb-forms do not normally appear w i t h an [ apo NP]PP

where the NP has the c h a r a c t e r i s t i c s of a prototypical agent, they
can neve rthe les s be considered true passives: all instances of
intransitive motion of this kind not c a r r i e d out by a p r o t o t y p i c a l
p a t i e n t are understandably expressed with the (b) f o r m s . 9 So, cases
of indirect causati on are expressed with t he active form
intransitive, e.g.:

(110) anevikame pano apo t i s fones tu (\l=aneveno)

we went u p ( s t a i r s ) ' f r o m' (because o f ) his screams

or the a c t i v e form t r a n s i t i v e - c a u s a t i v e , i.e .:

(111) mas a n e v a s e p a n o me t i s fones tu (V^anevazo)

he made us go u p s t a i r s / t o o k us u p s t a i r s with his screams

but not with the ' -ome' form. What is perhaps more interesting,
though, is t h a t even p r o t o t y p i c a l patients (obj ec ts not possessing a
s el f- movi ng mechanism, in this case) may wel l appear wi th the (b)
r a t h e r than the (c) forms, e . g . :

(112) 7 v a l i t s e s a ne vika n k j o l a s s to aeroplano [V ^aneveno)

the s ui tcases have gone up a lr eady onto the plane.

As a lr e ad y argued at the beginning of this section, the choice

between a ( di mi ni she d) passive and an a c t i v e form in transitive is
basically a ma tt er of degree o f 'agent-defocusing'. The degree o f
a g e n tiv ity of a p a r t i c u l a r predicate, even i f it is an instance o f a

- 187 -
genuine passive w i l l have to be decided i ndependently of the type o f
i mpl ied (but absent) agent. I t w ill r e s t w i t h a s p e c i f i c a t i o n of the
types o f obj ect s undergoing the motion described and the e xt ent to
which they may be shown ( i n specific sentences) to e xe rc is e control
over the event, have intention to carry it out or undergo it, in
short wi th a specification o f what have been e s t ab l is h ed as scaled
p r o p e r t i e s composing a g e n t i v i t y .

- 188 -
 Notes on Chapter 3

1. There i s , o f course, a d i f f e r e n c e between some n a tu r al forces at

least, and o t h e r kinds o f ' q u a s i - a g e n t s ' , in t h a t the former can
be understood as ' e v e n t s ' w h i l e the l a t t e r can be i n t e r p r e t e d as
r e f l e c t i n g extensions o f human motion. So, f o r i nstance:

(la) to s p i t i g r e m i s t i k e apo t o sismo { V=gremizome)

the house f e l l / w a s p u l l e d down from/because

o f the earthquake

wi th a mediopassive verb ' g r e m i z o m e f ( f a l l , be p u l l ed down) has a

parallel in ( l b ) :

(lb) to s p i t i e p e s e apo t o sismo (V = pe fto )

the house f e l l from/because o f the earthquake

w ith an i n t r a n s i t i v e verb ' p e f t o ' (fa ll). This p o s s i b i l i t y does

not arise in the case of ( 2 a) where the agent is human,
volitional, etc ., and can, therefore, only appear in an [ a p o
NP]PP related to a passive understanding of ' gremizome' (be
p u l l e d down):

(2a) to s p i t i g r e m i s t i k e apo t i n a stinom ia [V^gremizome)

the house was p ul l e d down by the p o l i c e

( 2 b) *to s p iti e p e s e apo t i n astinom ia (V=pefto)

the house f e l l by the p o l i c e .

It is th erefore suggested that ' sism os' ( e a r t hq ua ke ) , for

i ns tance, is in such cases understood as an event and can
f u nc t i o n as ' r e as o n' or stand f o r cause by i t s e l f , i . e . w ithout a
supporting environment, e . g . ' eks e t i a s t u s is m u ' (because o f the
earthquake).

- 189 -
2. I t must be shown t h a t aspect is not by i t s e l f the d e c i s i v e f a c t o r
here (as i t might be thought on the basis o f these two examples).
For:

( l a ) mia a s t r a p i b i e s x i s e t o n ura no ( V= 5 7 a s x i z o )
(a f l a s h o f ) l i g h t n i n g crossed the sky

does not have a corresponding

( l b ) *0 u r a n o s 5 ?a s x i s t i k e apo m i a a s t s r a p i (\= '6iasxizom e)

the sky was crossed by (a f l a s h o f ) l i g h t n i n g

but a much more emphatic d e s c r i p t i o n of a s i m i l a r s i t u a t i o n

(2) o lo k liro s 0 u r a n o s , S i a s x i s t i k e apo (\l=biasxizom e)

astrapes

the whole sky was crossed by

(flashes of) lightning

i s a cc ept abl e, presumably because i t is in t h i s l a t t e r case f e l t

that the sky is (perceptually) more seriously affected and
t h e r e f o r e c l o s e r to the p a t i e n t ' s p ro to t yp e . This expl ana ti on is
compatible wi th the distinctions made w i t h i n 'events' (in the
previous c hapt er ) between more and l es s typical instances of
'punctual o c cu rr e nc es ' . It can be thought t h a t i n (2) the event
has more d ur a t i o n ( i s l ess o f a punctual occurrence) than in ( l b )
and t h a t t h e r e i s , as a consequence, a l a s t i n g e f f e c t on the sky.
If something i s a f f e c t e d , it undergoes a change o f s t a t e and the
new s t a t e obvi ousl y has to l a s t long enough to be r e g i s t e r e d as
such. Appar entl y w h i l e in the case of ( l b ) t h i s p o s s i b i l i t y does
not a r i s e , in ( 2 ) , on the o t he r hand, i t does.

3. It is n a tu r a lly assumed t h a t m ani pul at iv e causati on i s r ul ed out

as a possibility for IC by d e f i n i t i o n . It is further assumed
that lexical caus at ive s w ill be r e a d i l y interpreted as d i r e c t /
m a n ip u l at i v e i f juxtaposed to e x p l i c i t ones; e . g . '5e/? t o n e s t i s a

- 190 -
 ey o o r Q i o , ton e k a n a a p l o s na s t a Q i ' ( I d i d not ' s t a n d ' him up, I
simply made him stand) or ' 5 en b o r o na t o n va lo mesa, boro omos

na t o n k a n o na b i ' (I cannot put him i n , but I can make him come

in). There also e x i s t s the p o s s i b i l i t y of a prototypical agent
s ubj ect combining wi th a prototypical patient object through an
exp licit causative: e.g . '/ x ris u la ekane tin porta na a n ik s i'
(Chryssoula made the door open) vs ' i x ris u la anikse tin p o rta '
(Chryssoula opened the door). The former example is a non-
prototypical description of the event and the implication is
probably t h a t the door would not open ( r e s i s t e d ope ni ng), so t h a t
excepti onal force or a speci al technique was needed to get it
open.

4. I am using ' s i m p le intention' as the second point on the

'vo litio n ' scale for all these actions/events for which only
context could provi de evidence f o r p a r t i c u l a r l y strong i n t e n t i o n .
This I consi der necessary in order to d i f f e r e n t i a t e between such
cases and those of verbs necessarily implying strong (or
'emphati c' as I c a l l it) intent, such as ' k a t a b i o k o ' ( chase).

5. I consi der that the theoretical assumptions made by the

proponents o f Prototype t he or y are compatible w i t h the p a r t i c u l a r
a p p l i c a t i o n o f a g e n t i v i t y c r i t e r i a in L i s t s IV and V. I t must be
pointed out, however, that t he re is no indication in the
literature o f whether or how such c r i t e r i a could be a ppl ied in
order t o compare l e x i c a l units (verbs i n t h i s case) w i t ho ut the
help o f a supporti ng c o n t e x t . Pluses and minuses do present an
idealizatio n , i n the sense t h a t a l l l i n g u i s t i c data are presented
in an i d e a l i z e d form when t hey appear in the form o f Tables and
Indices. All that can be said is that t he relevant data have
been at least checked a g ai ns t the intuitions of a number of
n a t i v e speakers. An x i s used t o i n d i c a t e marginal or r e s t r i c t e d
use, disagreement between subj ects and rather unpr ot ot yp ic al
uses.

- 191 -
6 . It could be thought t h a t the p o s s i b i l i t y discussed in connection
wi th ' p e rip la n je m e ' (wander) exists also for other agentive
verbs. Notice, however, t h a t (la) is very odd;

(la) ? p e r p a t u s a n apo laQos o b i y i e s (V = pe rp atao )

they walked because o f wrong i n s t r u c t i o n s

w h i le in ( l b ) :

( l b ) p e r p a t i s a n m e x r i t o staQmo apo l a Q o s o b i y i e s
they walked up to the s t a t i o n due t o wrong i n s t r u c t i o n s

the c r u c i a l PP ' a p o laQos o b i y i e s ' does not r e f e r to the a c t i v i t y

of wal ki ng but r a t h e r t o the goal 'to staQmo' .

7. Not ice that the verbs 'b java zom e ' (be read) and ' troyom e ' (be
eaten) correspond to a c t i v e form ones ' b j a v a z o ' (r e ad ) and ' t r o o '
(eat), respectively and t h a t in the Present tense they can be
only used as potentials. Examples of such c onst ruct ions
( p o t e n t i a l s ) are also o f f e r e d in 3 . 2 . 1 .

8 . A particularly illuminating r ece nt categorization of the whole

system of MG verbs on the basis of their s y n t ac t i co -s em a nt i c
f e a t u r e s , which attempts a break-away from the t r a d i t i o n a l active
vs passive voi ce dichotomy is presented i n Theophanopoulou-Kontou
(1 9 8 4 ) . Wi thi n this system, intransitives that we would
characterize as low in agentivity (on the basis of protot ype
theory criteria ), e.g. 'p eQeno' (die), are understood as
'unidirectionally patien t-oriented'. Verbs which can occur in
genui nel y passive c o n st r u ct i o n s, e.g. ' skotonome' ( be / g e t
k ille d ), w it h a corresponding genui nel y a c t i v e c o u n t e r p a r t ( i . e .
' skotono' (k ill)) are c at egor i ze d as ' m u l t i d i r e c t i o n a l ^ patient
and agent oriented' - the former l ab el attached t o the passive
i nstance and the l a t t e r t o the a c t i v e one. This approach can be
interpreted as also br in gi n g t og e t h e r non-agenti ve intransitives
wi th ( n on- a ge nt iv e) passi ves.

- 192 -
9. By ' p r o t o t y p i c a l patient' here I mean obj ect s not possessing a
sel f- movi ng mechanism or not using it w h i le the motion in
question is taking p l a ce . Theref or e animate subj ects are
compatible wi th the '-o/ne' form, provided they are being c a r r i e d ,
f o r i nst an ce .

- 193 -
 4. PRINCIPLES OF CATEGORIZATION AND
MINOR PROPERTIES OF MOTION VERBS

4 .1 P r i n c i p l e s o f c a t e g o r i z a t i o n o f motion verbs

In Chapters 2 and 3 a proposal was made as to how MGMVs can be

classified with r e spe ct to what are considered here 'major
classificatory features', i.e. properties relating to the S-P-E
distinction, causativity and agentivity. The content of such
properties w ill not be considered any further. Their s ta tus is,
however, worth di scussing as i t appears in d i f f e r e n t analyses of MVs
to be q u i t e d e ba t ab l e.

Two main poi nts w i l l be made in t h i s s e c t i o n . The f i r s t one involves

t he d i s t i n c t i o n between ' m a j o r ' and ' mi nor ' p r o p e r t i e s f o r MGMVs. It
is suggested here, t h a t the former e x h i b i t d i f f e r e n t c h a r a c t e r i s t i c s
from any o f the latte r kind o f p r o p e r t i e s which have been proposed
f o r the a n a l y s i s o f s i m i l a r semantic f i e l d s , e.g. English and German
motion verbs (and n a t u r a l l y those which w i l l be proposed as v a l i d f o r
MGMVs).

The second point i nv ol ve s the controversial issue of how such

semantic fields are or ga ni zed. Taxonomies for verbs and MGMVs in
particular are f i r s t discussed in d e t a i l . A separate di scussion o f
' mi nor ' properties proposed f o r MGMVs f o l l o w s in the next s ec t i o n .
The o v e r a l l organization of the field under investigation and the
'n o n-arb itra rin es s of categories' are examined in the last two
s ec t i o n s . It is also suggested that 'hierarchy of properties' is
different from 're la tiv e salience' and that tests e liciting
i nf or ma ti on from native speakers are needed to investigate both
relative salience and the psychol ogi cal validity of the proposed
types o f o r g a n i z a t i o n . (How t h i s i s to be e f f e c t e d i s the concern o f
the next c h a p t e r . )

- 194 -
What are p os it ed here as major p r o p e r t i e s f o r MGMVs have been shown
t o e x h i b i t t he f o l l o w i n g c h a r a c t e r i s t i c s . They are:

( a) C l a s s i f i c a t o r y , i.e. a p p l i c a b l e to l a r g e areas o f the vocabulary

and capable o f c a t e g o r i s i n g many d i f f e r e n t verbal
domains. The S-P-E d i s t i n c t i o n , causativity and
agentivity are evidently relevant for a
c l a s s i f i c a t i o n o f almost the whole o f the verbal
vocabul ary o f MG or Engl is h.

(b) Relative/graded, i . e . possessed by verbs to a h i gh e r or lower

degree. The fact that agentivity and
t r a n s i t i v i t y are graded p r o p e r t i e s i s recognized
in Hopper and Thompson (1980, 1982) and Givo'n
(1984). S-P-E i s also shown1 here as i n v o lv i n g a
continuum.

(c) I nvolving/subsuming ot he r fe atures, e.g. duration, i ng re ss i on ,

intention. The rel ev anc e of ' intentionality'
for an assessment of agentivity and of
'duration' for the S-P-E distinction has been
demonstrated in Chapters 3 and 2 r e s p e c t i v e l y .

(d) Characterising predications mainly (and verbs through

predications) and therefore linked to
grammatical categories, like tra n sitiv ity ,
aspect, e t c .

'Minor' properties relevant to the specific domain under

i n v e s t i g a t i o n do not e x h i b i t e i t h e r the f i r s t or the f o u r t h of these
characteristics. T h e i r r e l a t i o n t o (b) and ( c ) w i l l be shown t o vary
and depend on the na tur e o f each s p e c i f i c p r o p e r t y . This issue w i l l
be discussed i n s e c t i on 4 . 2 .

The d i s t i n c t i o n proposed here between major and minor p r o p e r t i e s has

a close p a r a l l e l in Putnam's (1975) d i s t i n c t i o n between s y n t a c t i c and

- 195 -
semantic markers, on the one hand, and s t e r e o t y p i c a l properties on
the other. Putnam's semantic markers are "category indicators of
high c e n t r a l i t y " and c o n s t i t u t e " pa r t o f a w i de ly used and important
system of classification" ( p . 2 67). An example of the sor t of
semantic markers Putnam uses f o r natur al kind terms i s o f f e r e d in 1.3
where the normal form d e s c r i p t i o n f o r 'w ater' i s presented. Natural
kind terms c l e a r l y i nv ol ve a completely d i f f e r e n t type o f l i n g u i s t i c
competence than motion verbs, so the parallel cannot be too c lose.
Besides, the l i n k between major c l a s s i f i c a t o r y p r o p e r t i e s f o r MVs and
grammatical c a t eg or i e s such as tra n s itiv ity and aspect makes it
p l a u s i b l e t o t h i n k o f them as s ynt a c t i c o- s e m a nt i c . So no d i s t i n c t i o n
parallel to Putnam's syntactic and semantic markers is envisaged.
They are also shown t o be graded ( u n l i k e semantic markers f o r natur al
kinds) and to subsume o the r f e a t u r e s which share c h a r a c t e r i s t i c s with
those p r o p e r t i e s which are considered here as minor p r o p e r t i e s .

A serious problem w i t h i n the lim its of the present study is the

s t at us of what may be considered 'in between' properties and in
particular those c h a r a c t e r i z i n g whole semantic domains. A case in
p oi n t is 'motion', which is n e ce s s a r i l y i mpl ied by a l l the verbs
included in the semantic field under investigation, and may be
thought o f as a hi ghe r l e v e l pr ope rty in r e l a t i o n t o s p e c i f i c and (by
im plication) minor f e a t u r e s o f motion. Adhering to the d i s t i n c t i o n s
s et out here between major and minor p r o p e r t i e s , ' moti on' appears as
a genui nel y 'in between' case. It may be thought of as
cl a s s i f i c a t o r y (although it is not of the same s ta tus as S-P-E,
a g e n t i v i t y and c a u s a t i v i t y ) since i t c h a r a c t e r i z e s a r e l a t i v e l y l a rg e
semantic field. Not ice that Pulman (1983:108) considers the
p o s s i b i l i t y o f i n c lu d i n g 'move' in the set o f ' p r i m i t i v e ' verbs such
as 'cause', ' make', 'become', 'act' and ' s a y ' . Some o f these verbs
correspond to what are posi ted here as major properties (namely
' cause' and ' a c t ' ) . Gradedness is not , however, a c h a r a c t e r i s t i c of
'motion', in the sense that all the verbs investigated here are
understood as possessing this feature and are not d i s t i n g u i s h e d as
more or l ess m o t i o n a l . * In t h i s r e spe ct , at least, the d i f f e r e n c e
between 'moti on' and S-P-E or causativity and a g e n t i v i t y is quite
d ra m at i c . ^ The question i s t h e r e f o r e r a i s e d whether a h i e ra r c h y o f

- 196 -
features can be e s t a b l i s h e d within the field under investigation.
This question r e q u i r e s c a r e f u l consideration.

A partial taxonomy o f MVs w it h corresponding h i e r a r c h i c a l f e a t u r e s is

attempted by Nida (1975:95-7) and is worth consi der ing in detail.
N ida 's example can be d i ag r a m m a t i c a l l y presented as f o l l o w s :

move

2 . e . g . come, go, a r r i v e , e.g. t r a v e l , t o u r , j ou rn e y,

enter, e x i t , advance, wander, rove, ...
retreat, ...

[manner]

3. wal k, run, c rawl , ride, hop, s k ip , ...

This example i s used by Nida t o i l l u s t r a t e the po i nt t h a t t h e r e are a

number of restrictions on semantic subor di nat ion : that one cannot
j ourney by hopping or s ki p pi n g, although walking is a possibility;
and t h a t although wandering i s u s a l l y done by w al ki ng, one can also
wander by t r a i n ( a t l e a s t according t o N i d a ) .

To f u l l y a p p r e c i a t e the problem w it h t h i s type o f approach, one must

also consider a further comment Nida makes (ibid.) concerning the
st atus o f 'tra v e l'. He argues t h a t since 'travel' does not spec if y
'd ire ctio n ', one could assume that it is more inclusive than the
terms 'come' and 'go'; yet there are many contexts in which these
l a t t e r terms are used in a more general sense than ' t r a v e l ' : "a term
such as ' go' may be used in such w i d el y d i f f e r e n t context s as 'go to
the di ni ng room' or 'go t o Europe' w h i l e ' t r a v e l ' would f i t only the
l a t t e r o f these contexts" (ib id .).

- 197 -
It is not in the l e a s t obvious why 'manner' s p e c i f y i n g MVs should be
c l a s s i f i e d as t h i r d l e v e l categories, i.e. lower than verbs which are
'general' in t h a t they do not s p e c i f y 'manner'. On the o th e r hand,
verbs specifying 'd ire c tio n ', even if t he y are s uper or di na te as
individual items to o th e r MVs (including some verbs specifying
'manner'), are not t o be understood as h i e r a r c h i c a l l y h i gh e r as a set
to the whole set of 'manner' specifying MVs. In other words, it
cannot be a p r i o r i e s t a b l i s h e d whether ' d i r e c t i o n ' i s more i n c l u s i v e
( g e n e r i c ) than 'ma nne r' .

If 'absence of certain features/specifications' were t o be equated

wi th ' r e l a t i v e l y hi ghe r l e v e l ' in MV-taxonomies, then we would expect
verbs l a c k i n g both a 'manner' and a ' d i r e c t i o n ' s p e c i f i c a t i o n to be
the most inclu sive ones and a l l o t he r MVs t o act as t h e i r genuine
hyponyms. In p r a c t i c e t h i s would mean t h a t the most i n c l u s i v e MVs in
this set would be 'wander' and 'rove'.^ In effect, it is obvious
t h a t the opposite holds. The i n t u i t i o n that 'tra v e l' and ' t o u r ' are
not really l es s specific than 'walk' and 'run' (as N id a ' s general
schema would have them) i s corrobor at ed by t he f a c t t h a t the former
are precisely t he ones o f f e r i n g fewer possibilities of combination
wi th 'manner' specifying MVs: They were touring crawling/hopping/
skipping are p a r t i c u l a r l y odd sentences, i f not unacceptable.

Although the terms appearing in N i d a ' s schema are shown not to be

t a x onom ica ll y or gani zed, o t he r terms within the field of MVs do
exhibit hyponymic relations. In what follows, it w ill be
demonstrated t h a t the whole f i e l d is n e i t h e r a p e r f e c t taxonomy, nor
a perfect paradigm, but has characteristics of both types of
organization. Specific subsets which c o n s t i t u t e taxonomies seem to
have p r o p e r t i e s quite d i f f e r e n t from those o f w e l l - e s t a b l i s h e d noun
taxonomies. It i s suggested, t h a t these p r o p e r t i e s are not s p e c i f i c
to the field of MVs, but are probably characteristic of verb
taxonomies in g e n e r a l . In the f o l l o w i n g s e c ti o n they are considered
i n some d e t a i l .

- 198 -
4.1.1 Taxonomies f o r verbs

The main poi nt s that need to be raised in d i s t i n g u i s h i n g noun

taxonomies ( o f the kind best known from e t h n o s c i e n t i f i c s t u d i e s ) from
verb taxonomies, involve: number of taxa, number of levels of
inclusiveness, cross-classification, c o ve r t categories, relative
abstraction of levels and the 'type of' or 'kind of' relationship
considered as characteristic of ta x onom ic al ly subor di nat e terms,
sometimes i n d i s t i n g u i s h a b l e from c l a s s - i n c l u s i o n ( Le h r e r 1 9 7 4: 2 0, 2 3)
but sometimes cont ra st ed to simple hyponymy (Rhodes 1 9 8 3 : 2 , 1 8 ) .

In its most general understanding, a taxonomy involves simply a set

o f words r e l a t e d through s upe ro r di n at io n t h a t can be repr esented as a
tree diagram, so t h a t any number o f words can appear at the same
level. Since, however, d if fe r e n t writers and d i f f e r e n t disciplines
have tended t o use 'taxonomy' t o imply d i f f e r e n t c o n s t r u c t s , i t seems
necessary to start this presentation wi th 'perfect'/genuine folk
taxonomies of the kind offered in Berlin, Breedlove and Raven
( h enc ef or th BBR) (1968, 1 97 3) , Pulman (1983) and Rhodes ( 1 9 8 3 ) . Such
s t r i c t taxonomies i n v o l v e :

1. Five l e v e l s o f i nclus ive ne ss a t most

2. Only one p os si bl e h i e r a r c h y o f taxons
3. No c r o s s - c l a s s i f i c a t i o n

The terms ' c a t e g o r y ' and ' t a x o n ' are used here i nte r ch a ng e ab ly and so
are ' w o r d ' , 'taxon-label', 'term '. The l a b e l s used f o r the d i f f e r e n t
l e v e l s o f i n cl us i ve n es s d i f f e r from one taxonomist t o t he o t h e r , so
in the example which f ol l ow s immediately (and which is p artially
reproduced from Pulman 1983:84), the ter mi nol ogy of BBR (1973),
Rhodes (1983) and Rosch (1977b) is juxtaposed for easy reference.
For the di scus si on, Rhodes' terms w ill be used, as being less
burdened w it h e t h n o b i o l o g i c a l i m p l i c a t i o n s than BBR's.

- 199 -
Level BBR Rhodes Rosch

( 0) PLANT unique inclusive

beginner

( 1) TREE FLOWER VEGETABLE life form kind super or di nate

(2) OAK WILLOW PINE generi c ge ner ic basic l e v e l

(3) DWARF WEEPING WHITE specific specific subordinate

(4) STRAIGHT- CURLY-LEAVED varietal varietal

(Fictional)

BBR (1973) and Rhodes (1983) contend t h a t the most i n c l u s i v e l e v e l is

of t en covert, and t h a t this may also be the case wi th a number of
t ax a a t o t he r l e v e l s . They also observe t h a t some g e ner i c taxa do
not belong to any kind and argue t h a t ge ner ic taxa have a c o g n i t i v e l y
privileged s ta t us (compare Rosch's ; basic level' categories). In
connection with the number of levels of i n c lus ive ne ss it w ill be
shown t h a t in most cases only two l e v e l s ( g e ne r ic and s p e c i f i c ) can
be fa irly easily established. Cross-classification is not only
frequent, but seems to be a characteristic of verb domains. The
depth of verb taxonomies and the lexicalization of c a t eg or i e s are
i n t e r r e l a t e d i ssues, and f urt hermore they l i n k to d i f f e r e n c e s between
levels in terms of relative abstraction. Nida (1975:89) observes
t h a t f i v e and s i x l e v e l s are not a t a l l rare in the case o f e n t i t i e s
(nominals) but t h a t "for events t h r e e and f o u r levels are not too
common unless we also use 'hi gh l e v e l meanings', e.g . e v e nt , action,
movement". It is rather obvious that even the commonest kind of
events are more complex than such p e r c e p t u a l l y and/ or functionally
distinct individuals as 'tre e s', 'dogs' and 'ch a irs ', commonly
appearing in noun-taxonomies. It is equally obvious that 'hi gh
level' units of t he t ype Nida mentions are also more complex than
noun- categor ies which correspond to th e m '(in terms of levels of

- 200 -
i n cl us i v e n e s s ) such as 'animal' or 'p la n t'. The questi on is not
whether one i s allowed t o use 'hi gh l e v e l meanings' but whether they
could p os si b l y l e x i c a l i z e in the way pl ants and a r t i f a c t s do.

As al re ady demonstrated, such high l e ve l properties are graded and

sometimes linked to specific p r e d i ca t i o n s and s p e c i f i c verb forms.
In the domain o f MGMVs, f o r i ns tan ce , 'states' can be onl y r e l a t e d to
P e r f e c t b' participles (as suggested in Chapter 2 ) . This means t h a t
one would have to e s t a b l i s h d i f f e r e n t taxonomies f o r d i f f e r e n t verb
forms, which is both in tuitively wrong and m e thodol ogi ca l ly
u nd e si r ab l e. Even if one concentrates on a particular verb form
(e.g. Pres. If.) and an agreement is reached to the effect that
'tre m o ' (tremble), for i ns tan ce , is a ' pr ocess' (as a lr eady
suggested), how does one l e x i c a l i z e 'pr oc ess ' to serve t h i s purpose?
For ' e v e n t s ' the usual substitute is DO. But DO, HAPPEN, CAUSE and
the l i k e , r e q u i r e a p r i o r agreement as to t h e i r exact content before
they can be used in any actual verb taxonomy. They are t e c hn i ca l
terms o f a compl etel y d i f f e r e n t nature from ' g e n e r i c ' and ' s p e c i f i c '
level categories such as 'walk' and ' s t r i d e ' , i.e. 'n a tu ra l', non­
technical lexical items. It is arguable that the inclusion of
t e c h ni c a l terms such as ' r e p t i l e ' or 'mammal' in noun taxonomies also
results in 'h y b rid ' taxonomies, i.e. taxonomies which are neither
pu r el y t e c h n i c a l nor pu r el y 'fo lk '. Not ic e, t h a t DO and HAPPEN are
even more a b s t r a c t and more t e c h ni ca l than t he above mentioned nouns.
They are no more examples o f l e x i c a l i z a t i o n than c o ve r t c at e gor i es
are. They are bundles of features/conditions/properties. Their
hierarchical or n o n - h i e r a r c h i c a l relations t o other f e a t u r e s can be
discussed on the basis o f Lounsbury's (19 64 :1 085 -7 ) p r i n c i p l e s which
w ill be d e a l t w it h in a subsequent s e c ti o n. The p o i n t is, however,
that they cannot acqui re the st at us of actual lexical items, i.e.
l a b e l l e d taxons w i t h i n a taxonomy.

It is therefore argued, that it is unlikely to have terms on more

than two or t h r e e levels of i nclusi veness i n verb domains, wi thout
i nc lu di n g technical terms and/ or cover t categories. Covert
c at e g o r i e s a lso raise a number of problems. The standard
taxonomists' j u s t i f i c a t i o n f o r s e t t i n g up c o ve r t c a t eg or i e s i s t h a t a
s er i e s of 'sorting tasks' performed by d i f f e r e n t groups o f subjects

- 201 -
on the same set o f terms - each group g e t t i n g d i f f e r e n t instructions
on the number o f c l u s t e r s they should form - can y i e l d hierarchical
structures, some l e v e l s o f which are u n l a b e l l e d . These unlabelled
taxons c o n s t i t u t e c ove rt c a t e g o r i e s . I f such a p r a c t i c e i s f ol l o w ed ,
the psychological reality o f c ove rt c a t eg or i e s cannot be questioned.
What can be questioned i s t h e i r exact s t a t u s . What we do know about
them is that the corresponding groupings of o bj e ct s are not
lexicalized . It seems th erefore plausible to assume, that c ove rt
c a t eg or i es are combinations o f p r o p e r t i e s , s i m i l a r t o t e c h n i c a l terms
and sharing t h e i r problems. For the moment, some c o ve r t c a t e g o r i e s ,
the content o f which can be s p el l e d out, are included in List VI,
which presents as many taxonomic relations as I have been able to
detect in t he area of MGMVs. In the di scussion of particular
t ax onomi ca ll y organized s e t s, it w ill become obvious that c ove rt
c at e go ri e s are not a l l o t t e d a s p e c i f i c p o s i t i o n in the h i e r a r c h y .

Verb taxonomies are unanimously recognized as involving fewer taxa

than noun taxonomies and t h i s makes i t a lr e ad y more d i f f i c u l t to see
if a particular fie ld i s t a x onom ica ll y organized or not (e.g. Rhodes
1 9 8 3: 1 0) . They also i n v ol v e fewer l e v e l s which renders i t impossible
to check c e r t a i n p r o p e r t i e s suggested as c h a r a c t e r i s t i c o f taxonomies
(e.g. maximum o f five levels). As can be testified in List VI,
c r o s s - c l a s s i f i c a t i o n i s abundant. This is excluded by BBR (1973) f o r
i nstance but allowed by more r e c e n t s t u d i e s , e.g. Hunn ( 1 9 8 2 ) , w hi le
Lehrer (1974) also allows f o r o v er l ap .

Taxonomic r e l a t i o n s which can be f a i r l y safely identified, i nvolve

'generic' and ' s p e c i f i c ' level categories. The o r i g i n a l idea on how
the inclusive term isto be identified is based, in the present
analysis, on Di xon's (1971) 'n uclear' verbs and the w e l l - e s t a b l i s h e d
procedure o f simple s u b s t i t u t i o n . Dixon (1971 :4 36) argues t h a t the
lexical verbs o f a language f a l l into two m u tu a l l y e x c l u s i v e set s:
nucl ear and non-nucl ear ones. The former type are anal ys abl e through
a small set of rather general and w e l 1 - mot ivated semantic f e at u r e s
(some o f which are l i k e l y to underlie c a t eg or i e s i n the grammar o f
the language) while the semantic content of the latter type is
definable through the semantic d e s c r i p t i o n of a n u c le ar verb ( or a
p r e vi o u s l y de fi ned non-nucl ear one) and the syntactic apparatus of

- 202 -
the language. One o f the c l e a r e s t examples Dixon o f f e r s involves the
verbs 'lo o k ' and 'stare': the former, which is nuclear, cannot be
defi ned through some ot he r verb, w hi le the latter, which is non­
n u c l ea r , can be de fi n ed as ' l o o k h a r d ' . One does not have t o agree
w it h the componential description Dixon assumes as the only one
a p p r op r ia t e for n uc lea r verbs in order to accept hi s main p oi n t s .
This d i v i s i o n i s a n a t u r a l one, in the s p e c i f i c sense t h a t i t is not
arbitrary (cor robor at ed by the evidence he brings in from Dyalyrjuy
and Guwal verbs). If a language has a minimum number o f verbs it
need not contai n any non-nucl ear ones, as i t can r e pl a c e them with a
'd e fin itio n '. For the same reason i t could not do w i t h o u t nucl ear
verbs (as they are not ' s u b s t i t u t a b l e ' ) .

Since i t has seemed important to include as many hyponyms as possi bl e

and given the speci al sociolinguistic problems of MG, a lr eady
mentioned in the Introduction, terms o f d i f f e r e n t origin and use as
well as verbs belonging to distinct registers are included.
'Katharevousa' terms o f r e s t r i c t e d use are marked w i th a s ubsc ri pt
SI while dialectal and 'lite ra ry ' ones are marked with S3. The
remaining terms are unmarked. They would c o n s t i t u t e an S2 category
which i s understood as the main body o f what i s commonly r e f e r r e d to
as 'Koine Nea El 1 i n i k i ' , i.e. common/standard Modern G ree k. 4 These
unmarked terms belong to different registers. So, for i ns tance,
among the verbs roughly equivalent to 'leave' or 'set o ff', both
'fe v y o ' and ' a n a x o r o ' are unmarked in L i s t VI (Category 5 ) . Besides
covering the same conceptual area these verbs belong to 'Koine Nea
Ell i n i k i ' , so choice between them depends on factors related to
d i f f e r e n t r e g i s t e r s or d i f f e r e n t ' fr a me s' ( i n the sense o f Brown and
Yule 1 983:238-241, or Verschueren 1 981:338). In very simple terms an
important person (e.g. a politician) can be said t o 'a n a x o ri' (set
off) to London, but the same verb cannot be used in a sentence
roughly e q u i v a l e n t t o 'my b ro t h e r l e f t f o r school at 8 : 3 0 ' . In such
a case ' f e v y o ' ( s e t o f f ) would be used i n s t e a d . Although an accurate
application o f these notions lies outsi de t he scope o f the present
study, it is necessary to refer to them, as the whole issue is
c l o s e l y r e l a t e d t o the di scussion o f basic l e v e l c a t e g o r i e s .

- 203 -
In connection wi th the p r e s e n t a t i o n o f the m a t e r i a l i n L i s t VI t hr ee
more poi nts must be made. There are separate c at e go ri e s (e.g. 6,7)
f o r corresponding t r a n s i t i v e and i n t r a n s i t i v e verbs. In those cases
where no cover term e x i s t s and the content o f the i n c l u s i v e c ove rt
category cannot e a s i l y be descri bed, sets o f verbs appear which are
understood as belonging t o the same na tur al class (he nc ef or th NC).
Cross-classification constitutes a distinct characteristic of the
taxonomies presented here (as a lr e ad y pointed out) and terms which
appear i n more than one category/taxonomy are marked w i t h CC followed
by the address o f the c a t e g o r y ( i e s ) i n which they a lso appear ( e . g .
'a n a rixo m e ' CC25 (climb)). There are many areas where class-
inclusion is only partial and some of these w ill be discussed
s e p a r a t e l y f o r each one o f the taxonomies which w i l l be analysed in
some d e t a i l in t he following section. A more acute problem which
arises and r e q u i r e s s peci al attention involves the d e t e rm in a ti o n of
the l e v e l o f i nc lus i ve ne s s o f the items under i n v e s t i g a t i o n .

4.1.2 Levels o f i nc lus i ve ne s s and l i n g u i s t i c a l l y unmarked c at e gor i es

As a lr ea dy mentioned in 4.1.1, taxonomists r e f e r to five levels of

inclusiveness, each one of which is attributed special
characteristics. In t h i s secti on a t t e n t i o n i s drawn to t he ge ner ic
and specific levels and their relation to Rosch's basic and
subordinate l e v e l categories.

In BBR ( 19 73:216) it is exp licitly stated that g e n er i c level

c a t eg or i e s are "the most commonly r e f e r r e d t o groupings o f organisms
in the na t ur a l environment, are the most s a l i e n t p s y c h o l o g i c a l l y and
are l i k e l y to be among the f i r s t taxa l earned by the c h i l d " . It is
fa irly obvious t h a t the ge ner ic level corresponds to Rosch's basic
level of abstraction which is the l e v e l at which " c a t e g o r i e s c a rr y
the most information, possess the hi ghest cue validity, and ar e,
thus, the most differentiated from one another" (Rosch et a l.
1 9 7 6 :3 83 ) .

- 204 -
The c o g n i t i v e l y privileged s ta tus o f g e n er i c l e v e l categories seems
to acqui re some substance through the experiments o f Rosch and her
colleagues in the case o f the kinds o f nouns they have i n v e s t i g a t e d .
In one such experiment, biological taxonomies were t e s t e d using as a
hypothesis the results o f ant hr opol ogi cal studi es (Berlin 1971, in
particular). Categori es like 'maple', 'b irc h ', ' o ak ' were expected
to belong to the basic l e v e l and ' t r e e ' (the i n c l u s i v e cat egor y) to
'k in d ' (see 4 .1 .1 ), i.e. Rosch's super or di na te level, while 'white
oak' and 'red oak', hypothesized as 'sp e cific' by Berlin should
correspond to Rosch's subordi nate l e v e l .

The test results showed that the hypothesized correspondence was

wrong. The basic level of abstraction identified by Rosch on the
basis o f the number o f common a t t r i b u t e s provided f o r each term by
subjects (see 1.3.2.3) turned out to be one level hi gher than
Berlin's hypothesized ge ner i c l e v e l , i.e. 'tre e ', proved t o be basic
level rather than 'oak' or 'b irch ' (Rosch 1977b:214-6). The same
a ppl ied to biological taxonomies for 'fis h ' and 'b ird ' also
hypothesized as s uper or di na te and proving t o be basi c l e v e l . These
results seem t o be in accordance w it h common sense and i n t u i t i o n and
it i s i n t e r e s t i n g to n o t i c e t h a t hypotheses based on the i n t u i t i o n o f
t he experimenters f o r n o n - b i o l o g i c a l taxonomies y i e l d e d the expected
results.

Unfortunately Rosch's method of identifying basic level c a t eg or i e s

presupposes the p o s s i b i l i t y of obtaining listings of attributes for
each term from s u b j ec t s. As a lr e ad y i n d i c a t e d in 1 . 3 . 2 . 4 this ta sk
cannot be implemented f o r verbal noti ons , t he l a t t e r being f a r more
abstract and complex than the sor ts of nouns used in Rosch's
experiments. It is therefore necessary to t ur n to ot he r related
notions i n or de r t o i d e n t i f y the l e v e l of inclusiveness/abstraction
of specific verbs as wel l as focus attention on the special
characteristics a t t r i b u t e d t o ge ner ic l e v e l categories. The notions
of 'cognitively privileged s ta tu s ', ' ps y chol ogi c al salience' and the
like, although intuitively correct are too vague to be used as a
working hypothesis. Rosch's basic level categories may be also
understood as being more f re que nt and f a m i l i a r than s uper or di na te or
s ubordi nate ones (Pulman 1 98 3 :1 25 ) .

- 205 -
Fam iliarity and commonness or frequency o f contact raise of course
t he quest ion: "fam iliar to whom?" I nt e rp er s on a l variation in how
basic a term is, is obvi ousl y to be expected in any semantic f i e l d .
It i s q u i t e concei vabl e t h a t Gricean p r i n c i p l e s are a t work governing
the choice o f taxonomic l e v e l from which a term i s drawn, at lea s t
for those models for which fam iliarity and sim ilar notions are
decisive in the choice of level of i n cl us i v e n e s s . Evidently t h er e
are also language and culture specific conventions, a gai nst which
such p r i n c i p l e s o pe r a t e . Now an a nal ys is o f the i n t e r p l a y o f f a c t o r s
such as assumptions on knowledge o f speaker and l i s t e n e r , cont ext ,
situation, etc. lies outsi de the scope of the pr esent study.
N e ver the le ss , the i m p l i c a t i o n s o f such consi de ra ti ons f o r determining
the l e v e l o f i n cl us i ve n es s o f verbs have to be poi nted out.

Consider as an example a subset o f category 5 in L i s t VI i n vol vi ng

verbs a l l o f which can be used f o r a ship s a i l i n g o f f :

'a p o p le o ', ' anaxoro', ' fe v y o ', ' s a lp a r o '.

As a l re a dy mentioned, 'a n a x o ro ' and 'fe v y o ' are not restricted to

sailing (off) and belong t o d i f f e r e n t r e g i s t e r s . Sim ilarly 'a p o p le o '

is formal and may be considered te chnical. Notice that r a d io and

television broadcasts may use 'a p o p le o ' but are more l i k e l y t o use
'a n a x o ro ' i n st ea d. The only term which is s o c i o l i n g u i s t i c a l l y marked
here (S3) is 's a lp a ro ', considered somehow d i a l e c t a l , very ' p o p u l a r '
and t h e r e f o r e perhaps not p a r t o f 'Koine Nea E l l i n i k i ' (although i t
i s not uncommon in novels or t r a n s l a t i o n s o f f i l m - s c r i p t s ) .

On i n t u i t i v e grounds, at least, t he re can be l i t t l e doubt t h a t for

specialists or non-specialists, educated or non-educated native
speakers the most 'cognitively privileged', 'psychologically
s a lie n t', f re que nt and f a m i l i a r verb i n the s et i s a c t u a l l y 'f e v y o ',
which i s also the one f i r s t l earned by the average c h i l d . What are
the consequences of these observati ons in terms of levels of
i n c lu s io n? Following Dixon ( 1 9 7 1 ) , since ' a n a x o r o ' and ' f e v y o ' cover
the same conceptual ground and the remaining verbs can be 'defined'
by referring to them, they should appear on t he same level.

- 206 -
Following Verschueren ( 1 9 8 1 ) , d i f f e r e n c e s i n r e g i s t e r are accountable
f o r in terms o f frames and t h e r e f o r e each one o f these verbs is basic
level for a d iffe re n t subset o f the popul at ion: 'a p o pleo ' f o r navy
officers, ' s a lp a r o ' for sailors, 'anaxoro' f o r educated Greeks wi th
a strong tendency towards pomposity or television broadcasting
personnel at work, 'fe v y o ' for people who have never travelled by
ship and more i m p o r t a n t l y f o r anybody who has no reason t o make any
of the discriminations i mpl ie d by and conveyed through the ot he r
verbs. In this sense it could be argued that 'fe v y o ' alone is
' l i n g u i s t i c a l l y unmarked'.

Words used in normal everyday speech, which do not require speci al

context ual features to be used appropriately, and c a rr y no
implications or affective overtones are termed 'lin g u istic a lly
unmarked' in Cruse (1977). The ne utr al quality of such words is
expl ai ned by Cruse on the basis o f well-known Gricean p r i n c i p l e s and
R. Brown's noti on o f a l e v e l o f 'maximum u t i l i t y ' ( i b i d . :155). Cruse
considers the relation between ta xonomi ca ll y linked words such as
'anim al-dog-spaniel' and observes that it is more o ft e n important
t h a t a spaniel belongs t o the class o f dogs than t h a t it belongs to
the class o f s p a n i e l s . In t h i s sense 'dog' r epresents the 'neutral
level of s p e c ific ity ' (ibid.) and t h i s ne ut ra l q u a l i t y or l i n g u i s t i c
unmarkedness is said to be i nh e re nt in c e r t a i n items. Such items
belong probably to Rosch's basic l e v e l of abstraction.

In a given c ont ext a verb such as ' a popleo' or 's a lp a ro ' (both
equivalent to 'sail away') w ill be the normal way to descr ibe the
same event and can t h e r e f o r e be understood as a basic l e v e l term in
th a t p a r t i c u l a r context. This does not mean, however, t h a t 'a p o p le o '

and 's a lp a ro ' are not subordinate to 'fe v y o '. A taxonomy can be
e s t a b l i s h e d i ndependentl y o f which verbs are most normal, f a m i l i a r or
fr e qu e nt in particular c on te xt s . The obvious requirement is, that
the term posi ted as hi ghe r l e v e l than s p e c i f i c or subor di nat e, should
have the wel l known c h a r a c t e r i s t i c s (from t r a d i t i o n a l structuralism)
o f covering a wi der conceptual area than those terms pos it ed as lower
level ones, in which all the latter are i ncluded. If the
hypothesized subordinates are used correctly (i.e. appropriately),
their i n c lu s i o n relations can be assumed to be part of native

- 207 -
speakers' competence. Provided the term p o s it e d as i n c l u s i v e c a r r i e s
the additional characteristics of being lingu istically unmarked,
frequent, fam iliar and i n d i s p u t a b l y pr imar y, i t can be considered as
an i nstance of Di xon' s (1971) 'nuclear' ver bs, Rosch's basic l e v e l
categories, or the g e ner i c level t ax a of traditional taxonomists.
These (additional) properties may also c h a r a c t e r i z e verbs appearing
as subordinate in the taxonomy*, the difference is, of course, that
these l a t t e r verbs are not i n c l u s i v e .

The most important issue in the l i g h t o f Prototype t he o ry remains of

course the obs er vati on that subordinate c at e go r ie s are not
e q u i d i s t a n t from t he i n c l u s i v e one. Some o f them are judged as more
c h a r a c t e r i s t i c o f t he i n c l u s i v e category name than o t h e r s . We cannot
know, a priori, in the case of verbs, what renders subordinate
c a t eg or i es more or l ess prototypical o f the category name: Rosch's
method, which is again based on counting attributes provided by
s u b j ec t s, is inapplicable. T he r ef o r e , commenting on the content of
the various taxonomies, proposed here for MGMVs, factors possi bl y
contributing to ( r e l a t i v e ) p r o to t y p i c a li ty w i l l be poi nted out and in
particular: ( a) linguistic markedness and (b) relative class
inclusion. For unlike biological categories, verb taxonomies may
i nv ol ve partial inclusion. Besides these, relative salience rather
than number o f common f e a t u r e s seems to be o f g r e a t importance, an
issue which w i l l be discussed in the f o l l o w i n g s e c t i o n , d e a l i n g wi th
'mi nor p r o p e r t i e s ' and e s p e c i a l l y after a presentation of relevant
t e s t r e s u l t s ( i n the next c h a p t e r ) .

In order to keep this preliminary di scussion to what is only

necessary f o r the p r e s e n t a t i o n o f the taxonomies, a number o f poi nts
simply mentioned so f a r w i l l be taken up again and discussed in more
detail in the a n a l y s i s o f the s p e c i f i c m a t e r i a l examined her e. List
VI contains both the taxonomic sets and the n a tu ra l classes proposed
f o r MGMVs, but only those taxonomies which seem t o e xe mpl i fy the most
acute problems a r i s i n g here are discussed in d e t a i l i n what f o l l o w s .

- 208 -
4.1.3 Taxonomic sets proposed f o r Modern Greek motion verbs

As a lr e ad y mentioned, 'move' could be a candi date f o r an i n t e r m e d i a t e

level in a general taxonomy o f MVs. I t must be poi nted out here t h a t
the corresponding lexical!zed taxon in MG, namely 'kunjem e'
( mo v ej ^ j p) , can h a r d l y be understood as a d i r e c t inclusive term of
c h a n g e - o f - l o c a t i o n ver bs, as i t i s normally used t o imply change-of-
position/partial motion. The ref or e intransitive 'kunjem e' and
transitive 'k u n a o ' appear in d i f f e r e n t taxonomies (1 and 2 ) i n v ol v i ng
verbs which imply motion but no c h a g e - o f - l o c a t i o n .

The onl y possi bl e l e x i c a l i z e d super or di nate for a number o f change-

o f - l o c a t i o n MGMVs i s 'p iy e n o ' (go) which i s not , however, equivalent
to the English super or di na te TRAVEL ( f o r the s ta tus o f which t h e r e
also exist problems as it differs from t he common English verb
'tra v e l'). The f i r s t problem w it h 'p iy e n o ' (go) - s et 3 in L i s t VI -
is that most of the terms that could be posi ted as its hyponyms,
could be also hyponyms o f ' e r x o m e ' (come). In ot he r words, though i t
looks as the most general verb in the set under c o n s i d e r a t i o n , it
does not have t h a t status, because it contai ns a deictic component
which i s not present in the v as t m a j o r i t y o f the remaining verbs, A
further problem emerges if one a p pl ie s the standard substitution
te s ts : 'p i y e n o j ' (go) requires a goal specification, which is not
necessary f o r most o f i t s p os si bl e hyponyms. I f both ' p i y e n o j ' (go)
and ' e r x o m e ' (come) are put a t an i n c l u s i v e l e v e l , two p o s si b l e sets
o f hyponyms could appear as candidates f o r t he hypothesized taxonomy.
The f i r s t one i s i ncluded in L i s t V I :

- 209 -
 p iy e n o j / e r x o me

go / come

metaveno$} taksiSevo / petjerne2 d5evo$j porevome

go from NP|_qc travel / go f o r a\^ go towards go a long

t0 NPL0C short time pL0C d i s t an c e on
(dash, (coveri ng a foot
run out t o ) long d i s t a n c e )

proxoro opisQoxoro proserxom e^j k o p j a z o $3

advance retreat come/go come/go

Fi g. 1

Of these hyponyms onl y 'm etaveno' corresponds to 'p iy e n o ' and not to
7e r x o m e w h i l e 'p rose rxo m e ' and 7k o p j a z o ' are p r o p e r l y included in
' erxome' o nl y. The remaining verbs could be hyponyms of either
supposed s u pe ro r d i n a t e , a situation not like ly to appear in noun
domains.

In terms o f c o l l o c a t i o n a l patterns, notice t h a t 'p o re v o m e ', 7o5 e v o 7,

'p ro x o ro ' and 'o p is Q o x o ro ' do not n e c e s s i t a t e a goal specification;
if one does appear, it is i ntroduced w it h a d i r e c t i o n a l preposition
equivalent to 'towards', namely 'p ro s '. The same situation is
e v i de n t if a different set of verbs is posi ted under this 'twin'
inclu sive c at e go ry , namely 'manner' s p ec i f y i ng v er bs, such as:
'p e rp a ta o ' (walk), 's e rn o m e ' (crawl), 'k ila o j' (ro ll), 'p e ta o j'

(fly ), 'y l i s t r a o ' (slip , slide). Some o f the terms of both sets
appear as i n c l u s i v e t o ot he r c a t e g o r i e s (e.g. 'p e r p a t a o ' (walk), or
'p ro x o ro ' (advance)), and what i s more impor tant, a number of them
appear t o possess the c h a r a c t e r i s t i c s suggested here as r e l e v a n t to
basic level categories. This means that 'p iy e n o j' could be a

- 210 -
candidate f o r the f i r s t l e v e l o f inclusi veness ( l i f e - f o r m , i.e. kin d/
superordinate), its immediate hyponyms (of both s et s) could be
'g eneric' level and t h e i r own hyponyms would be ' s p e c i f i c ' level.

N ot i c e , however, the compl icati ons arising for a massive taxonomic

organization of this t y pe . It could be claimed that 'manner'
specifying verbs might also be considered as hyponyms o f 'general'
verbs , such as those appearing in F i g . l and many others such as:
'aneveno' (ascend), 'a ko lu Q o ' (follow ), 'b e n o ' (enter), in short any
motion verb which i s not marked f o r ' manner'. I f such a s o l u t i o n is
adopted, 'p iy e n o ' (go) would be l e v e l 1 (kind), all verbs which do
not s p e c i f y 'manner' would be l e v e l 2 (generic/basic), 'manner' verbs
would be l e v e l 3 (subordinate/specific) and t h e i r hyponyms would be
level 4 (v a rie ta l). The u n d e s i r a b i l i t y of such a s o l u t i o n i s a lr ea dy
noted in connection t o N i da ' s proposal which i s s i m i l a r , in a way, to
this schema - see 4.1. Such a schema would wrongly predict that
sentences such as: '? ? ta ksib e ve sernomenos' (he t r a v e l l e d c r aw l in g)
are acceptable (which they c e r t a i n l y are n o t ) . It would also imply
th a t ' p e r p a t a o ' (walk) i s lower l e v e l than h i g h l y s p e c i f i c terms such
as ' t r i y i r i z o ' (roam around). What is worse, 'p e rp a ta o ' (wal k) might
even have to appear at a lower level (i.e . 4=varietal), since
'p ro x o ro ' (advance), for i n s t an c e, has i t s own set o f hyponyms (see
35 in L i s t V I ) , which would have to appear on l e v e l 3 (generic/basic
level). Such p r e d i c t i o n s are i n t u i t i v e l y fe lt to be wrong and both
collocational and o t h e r s u b s t i t u t i o n consi de ra ti ons would have t o be
abandoned in more cases than not .

In short, although the s ta t us of 'p iy e n o j' (go) is certainly

different from t h a t of most other change-of-location verbs, it is
more plausible to consi der both 'p iy e n o ' and 'e rxom e' appearing
immediately under a c ove rt category i n vo l v i n g 'change-of-location',
and F i g . l sub-taxonomy as i n c lu d i n g a l l their ( more/ less c l e a r cases
of) hyponyms. If simple s u b s t i t u t i o n i s the c r u c i a l criterion, then
'erxom e' (come) has t o appear on the same l e v e l with 'p iy e n o j'. It
is, however, f e l t t h a t in t h i s d e i c t i c p a i r , 'p iy e n o j' i s somehow the
l e s s , marked member. The validity of such judgments can only be
checked through relevant tests, some o f which w i l l be discussed in
the next c ha pt e r.

- 211 -
The taxonomic set headed by 'p e rp a ta o ' (wal k) and 'v a b iz o ' (wal k)
needs to be examined in some d e t a i l as i t contains a l o t o f hyponyms
to these inclu sive terms and provides examples of most o f the
problems discussed from the theoretical po i nt of view in the
preceding s e c ti o ns .

perpatao / vabizo

[walk here and t h e r e ] [wal k / u n s t e a d i l y ] [walk t o g e t h e r w i t h ]

s aunt er , stroll

s u la tsa ro trik liz o fstra vop ata o \ s ino bip oro simporevome

stagger// stumble, \ \ (+ long d i s t an c e)

v o 1t a r o t r i p over

serjan parapatao (stumble, walk u n s t e a d i l y )

[wal k a long d i s t a n c e ] [walk slo wl y]

pezoporo I obiporo porevome vra b ip o ro

T
a ryo pa to ^ aryoporo
( + ' long d i s t an c e )

b u s u liz o / a r k u b i z o $3 tsa la vu to s i y a n o p e r p a t o $3
crawl (as o f a baby) wade wal k s l o w l y / w i t h o u t noise

v im tiz o
/ /
k i k l o b r o m o $2
i
b ra ske lizo p is o p a to 53 akropato$ 3

pace walk in c i r c l e s stride wal k backwards tiptoe

Fi g. 2

To s t a r t w i t h , since both ' p e r p a t a o ' and ' v a b i z o ' cover the same
conceptual a rea, they are t e n t a t i v e l y put on the same l e v e l

- 212 -
(hypothesized g e n e r i c / b a s i c l e v e l ) . For some speakers, a t l e a s t ,
' p e r p a t a o ' and 'v a b iz o ' are thought o f as belonging t o d i f f e r e n t
r e g i s t e r s and in t h a t case ' p e r p a t a o ' i s the unmarked member o f the
pair. I f such c o n si d e ra t i o ns are allowed t o pl ay a r o l e in the
o r g a n i z a t i o n o f t he taxonomy, ' v a b i z o ' could appear as a hyponym o f
'p e rp a ta o '. The cost i s not very g r e a t . It i s q u i t e possi bl e t h a t
'v a b iz o ' i s the most c h a r a c t e r i s t i c kind o f 'p e rp a ta o ', i.e . i t s most
prototypical instance. Therefore, i t may be considered e q u a l l y basic
l e v e l wi th 'p e rp a ta o ' and move one l e v e l up . 5

Considering now t he st at us of the proposed covert c at e go r ie s in

F i g . 2, n o t i c e f i r s t t h a t : 's in o b ip o ro ' (walk a long d i s t a n c e t og e t h e r
with) and i t s near-synonym ' s i m p o r e v o m e ' may be also c l a s s i f i e d under
the next c ove rt c at e go r y, namely [walk a long d i s t a n c e ] , or appear as
hyponyms of the verbs of this latter category which are
m o r ph ol o gi ca l ly related to them. Notice also that if c ove rt
c a t eg or i e s are p a r t o f the h i e r a r c h y , the b l a t a n t l y wrong p r e d i c t i o n
is made t h a t verbs such as ' t r i k l i z o ' ( s t a g ge r ) or ' p o r e v o m e ' (walk a
long d i s t a n c e ) are on a lower l e v e l o f i nclus ive ne ss than 'p is o p a to '

(walk backwards) or 'k ik lo b ro m o ' (walk in circ le s), simply because

'trik liz o ' ( st a g g e r ) does not happen to belong together wi th any
o t h er t er m( s) o f t he taxonomy.

Covert c a t eg or i e s are useful f o r grouping t og e t h e r items which are on

the same taxonomic level and which have an obvious common
characteristic, i.e. words which form a n a t u r a l class but e x h i b i t no
taxonomic r e l a t i o n s . Wi thi n such na tur al classes c e r t a i n items may
be l i n g u i s t i c a l l y unmarked w i t h r espect t o the remaining members o f
the same c l a s s . In F i g . 2 very few such cases can be identified:
'trik liz o ' (s tagger), 'p a ra p a ta o ' (stumble) and 'porevom e' (walk a
long d i s t a n c e ) . The same, however, a ppl ie s t o items which are not
under a c ove rt cat egor y.

Not ice t h a t , ' vim a tiz o ' ( p a ce ) , 'b ra s k e liz o ' (stride) and 'b u s u liz o '

(crawl (as o f a baby)) are also l i n g u i s t i c a l l y unmarked compared t o

o th e r hyponyms of 'p e rp a ta o ', such as 'p e zop oro', 'o b ip o ro ',

' v ra b ip o ro ' which are high r e g i s t e r , and the items marked SI or S3

- 213 -
which are not part of most speakers' a c t i v e vocabul ar y. Al l the
suggested hyponyms o f ' p e r p a t a o ' a r e, however, on the same taxonomic
level. The e x t e n t t o which t h e i r being more marked l i n g u i s t i c a l l y
affects their di st an c e from the i n c l u s i v e category can be discussed
only a f t e r p r o t o t y p i c a l i t y t e s t r e s u l t s are obtai ned.

No t i c e , fin a lly , that many items of this s et participate in ot he r

taxonomies as w e l l . The poi nts o f c r o s s - c l a s s i f i c a t i o n are marked in
List VI. To mention a few, notice that 'porevom e' (wal k a long
d i s t an c e ) is also p a r t o f the previous taxonomy ( F i g . l ) ; 'b u s u li z o '
and ' a r k u b i z o ' both implying 'crawl (as of a baby)' can also appear
together with 'sernom e' (crawl) in set 25; 'p is o p a to ' (walk
backwards) appears also in a taxonomy i n v o lv i n g backward motion in
g e n er a l , i.e. under ' o p i s Q o x o r o ' (move backwards); the s et o f verbs
implying 'walk here and t h e r e ' appear t o g e t h e r wi th 'triy iriz o ' (go
here and t h e r e , roam around) in set 18, e t c .

Another large taxonomy, which i ncludes a lot of hyponyms to the

hypothesized inclusive term and allows a number of i n te r m ed i a t e
cove rt c a t e g o r i e s , i s headed by 'fe v y o ' and 'anaxoro' both o f which
are equivalent to 'leave or set o ff'. It is suggested that only
'fe v y o ' (leave) in t h i s taxonomy represents t he l e v e l of s p e c i f i c i t y
which is "least moti vated contextually" in the sense of Cruse
(1977:156), and that 'a n a x o ro ' (depart) is lin g u istically marked
although i t covers a s i m i l a r conceptual area wi th i t (as pointed out
in the preceding section). The ref or e onl y 'fe v y o ' (leave) is
understood here as a basic l e v e l cat egor y.

- 214 -
 fevyo / anaxoro

leave / depart

[change country o f [ i n c re as e d i s t an c e] [ l e a v e / g e t o f f the

r esi dence] mo^e I wVAf rom gr ound/ land]
emi gr a te / im mi gr a te

m etanastevo \lksenitevom e< apom ne apoyio now e apopleo

move away \from t ake o f f \ sail off
a po Si mo im sevo 53 (t he ground)

m etiko $j ekpatrizom e param en zoa

\ \ \ v
move a s i d e / o v e r sa lp aro ,
sail off
ksemakreno ifjotravjem e
recede i n t o move away
the d i st an ce

[ l e av e as u n de s ir a bl e ] [ l e a v e s e c r e t l y / a l t h o u g h not allowed]
clear o ff, buzz o f f

/csekub 7 7a 0 5 3 Siafevyo \ ksefevyo \ brapetevo

get away ^scape, s l i p \ escape
s t r i v o S3 s p a z o S3 ... away

apoxoro aposirome aperxome3 j

withdraw retire l eave

Fig. 3

- 215 -
It is further suggested t h a t some o f the subordi nate c a t eg or i e s in
this taxonomy are a lso lin g u istically unmarked compared to ot he r
items which cover a s i m i l a r , if not i d e n t i c a l , conceptual area wi th
them. This means t h a t 'm e ta n a s te v o ' (e mi gr at e ) is a lin g u istically
unmarked term compared to th e remaining verbs i mplying 'change
country o f r e s i d e n c e ' , because the l a t t e r c a r r y a f f e c t i v e overtones,
are l es s fam iliar, belong t o marked r e g i s t e r s etc. This cannot be
i n t e r p r e t e d as i mplying t h a t i t should be on the same taxonomic l e v e l
with 'fe v y o ' (leave). It may be, however, an example of a taxon
moving up and l a b e l l i n g an otherwi se u nl a b e l l e d hi ghe r taxon ( i . e . a
c over t c a t e g o r y ) . This p o s s i b i l i t y i s o f f e r e d in Hunn (1982) wi thout
an accompanying specification of the circumstances under which a
taxon can g e n e r a l i z e in t h i s way. Wi thi n the framework adopted here,
the p o s s i b i l i t y o f a subordinate c a t eg or y ' s becoming basi c l e v e l is
restricted to c at e g o r i e s appearing as most p r o t o t y p i c a l of the
inclusive category name, and c ove rt c at egor i es are shown (for
independent reasons) to constitute no taxonomic level. In this
taxonomy, both 'a p o x o ro ' (wi thdraw) and 'a p o s iro m e ' (retire),
although r e l a t i v e l y l i n g u i s t i c a l l y marked (high r e g i s t e r ) seem to me
more prototypical of the inclusive category ('fe v y o ') than
'm e tan astevo ' (e mi gr at e ) or 'S ra p e te v o ' ( es cape), for reasons to be
discussed in the f o l l o w i n g c hapt er . As they are not l i n g u i s t i c a l l y
unmarked, however, the chances t h a t either of them g e n e r a l i z e s and
appears on the same l e v e l wi th 'fe v y o ' are few; but 'm e tan astevo '

and 'S ra p e te v o ' cannot be allowed to become basic level either, in

t h i s p a r t i c u l a r taxonomy, as they are most u n l i k e l y to be considered
very p r o t o t y p i c a l i nstances o f ' f e v y o ' .

At this stage some of the factors which may pl ay a role in the

formati on o f pr otot ypes in the area o f MGMVs are simply pointed out
and intuitive judgments are made which may not correspond to test
results. The i mportant t h i ng to notice is that since counting
attributes is not feasible, we are le ft with the vague notions of
perceptual and soci al s al i e n c e (Rosch and Mervis 1 97 5 : 5 9 9 ) . An
attempt i s made a t making these notions more e x p l i c i t and t h i s can
only be done in connection wi th very s p e c i f i c (and consequently very
restricted) da ta here. The g r e a t e s t , disadvantage i s the following:
except for v er y few cases, it is virtu ally i mpossible to find

- 216 -
c a t e go r i e s with a sufficient number of genuine hyponyms (properly
included s ubor di nat es) and e q u a l l y impossible to f i n d enough hyponyms
unmarked f o r r e g i s t e r , so t h a t these two f a c t o r s could be constant
and one would t h e r e f o r e have t o consider onl y r e l a t i v e salience of
a t t r i b u t e s or dimensions. A case in poi nt i s the f o l l o w i n g taxonomy.

'p e fto ' (fall)

[fall down, [fall on s . o . / s . t h . [fall into mass/liquid]

i . e due to g r a v i t y ] wi th f o r c e ]

k a t o l i s Q e n o 3 2l x a tra k ila o p rospefti nxnome vutaoj katabiome

s l i d e down roll down ge t on j throw di ve dive

one's knees \ 6 n e s e l f upon

ka ta p ip to $ i k u tru va la o ximao enskipto$ i

fall down* roll down charge

violently

sorjazome katareo or mao eform o$j

c o ll a p s e c o ll a p s e dash, f a l l v i o l e n t l y on

gremizome tubaro e p ip ip to $ j xinome

fall from fall upside dash, f a l l v i o l e n t l y on

a precipice/ down, overt ur n
fall in rui ns

anatrepom e

fall upside down,

ov er tur n

Fi g. 4

- 217 -
Three n a t ur a l classes can be identified here, two of them
differentiated in terms of the cause of motion and a third one
involving downward motion into liquid . As a lr e a d y suggested in
Chapter 3, 'p e fto ' (fa ll) is p rim arily (p r o t o ty p ic a lly ) understood as
implying downward motion due to gravity (hence it is low on the
agentivity scale). The firs t cover t category contai ns therefore
hyponyms which are as a set most c h a r a c t e r i s t i c of 'p e f t o ' (fa ll),
i.e. i nstances of 'fa ll down'. The second one i nv ol ve s verbs
implying 'fa ll on someone/something on purpose' and therefore
correspond to a f a i r l y unp ro to t yp i ca l understanding o f ' p e f t o ' , w hi le
the last set contai ns only two items, which can be r eplaced by
'p e fto ' only when the motion is into liquid and in p a r t i c u l a r into
the sea, namely 'v u ta o j' (dive) and its near-synonym 'ka ta bio m e ' .

These two items form a n a tu r a l class with others not implying

'fa llin g ' but 'sinking' and appear wi th them under 19 in List VI.
This taxonomy is th erefore the clearest case of relative
p r o t o t y p i c a l i t y o f whole sets o f items with r espect t o the i n c l u s i v e
basic level one. It seems fa ir to say that although all the
suggested hyponyms are p r op e r l y included in the hi gher one, those
implying fallin g down a c c i d e n t a l l y and wi thout rolling/turning are
the c lo s e s t to it (all o t he r t hi ngs being equal). These are
'g re m izo m e ' (fall from a p r e c i p i c e / i n r u i n s ) , 's o rja z o m e ' (collapse),
'k a ta re o ' (collapse), 'k a ta p ip to ' (fall down) and 'k a to l isBeno'

(slide); the last three verbs are l i n g u i s t i c a l l y marked. These are

fol lowed by items involving rolling/turning (t he remaining items in
the same c ove rt c at egor y) and f i n a l l y by those r e q u i r i n g a special
environment ( w a t e r ) . This may gi ve an idea o f r e l a t i v e d i s t an c e from
the inclusive cat egory in terms o f a t t r i b u t e s and r e l a t i v e s al i e n ce
of a ttrib u te s , the importance of which can be only discussed on the
basis o f t e s t r e s u l t s in the f o l l o w i n g chapt er .

The remaining sets are poor examples o f taxonomic o r g a n i z a t i o n , in

the sense t h a t very few items are pr op e rl y included i n the suggested
h i g he r l e v e l one. Three more examples w i l l be considered, starting
w it h ' viQ izom e' ( s i n kj N TR )-

- 218 -
 viQizome

sink

v u lja katapondizom e > emvaptizome3 j katabiom e

sink sink t o tfTe bottom submerge dive, submerge

vutaoj navayo fudaro

di ve founder sink t o the bottom

Fig-5

As can be a t t e s t e d i n L i s t V pr ese nt ing the d i f f e r e n c e s in a g e n t i v i t y

of var ious non-c aus ati ve MGMVs, 'viQ izo m e ' (s ink) involves
essentially a cc i de n ta l motion (due t o gravity) and only marginally
does it also allow for intended submersion into water. The ref or e
' katabiom e' (dive) and 'v u ta o j' (dive) are fa irly marginal in this
taxonomy as they n e c e s s a r i l y i n vo lv e i n t e n t i o n . Not ic e t h a t n e i t h e r
can be r epl aced by ' viQ izom e' unlike the situation in the previous
taxonomy where 'p e fto ' (fall) can r e pl ac e them both, presumably
because 'p e fto ' is l ess marked f o r absence o f i n t e n t i o n and contr ol
than ' v i Q i z o m e ' , a f a c t which does not show in L i s t V.

A f u r t h e r problem in F i g . 5 i s t h a t 'v u lja z o ' (sink) i s a near-synonym

o f the i n c l u s i v e cat egor y name and c e r t a i n l y l i n g u i s t i c a l l y unmarked.
The only reason why 'v iQ iz o m e ' appears as t he basi c l e v e l term in
F i g . 5 is that 'v u lja z o ' does not i n vol ve intentionality or contr ol
even m a r g i n a l l y (see L i s t V ) . I f we concentr ate on the p r o t o t y p i c a l
understanding o f ' v i Q i z o m e ' , both verbs should appear as i n c l u s i v e o f
the rest. A lternatively, 'v u lja z o ' w ill predictably be considered
the most c h a r a c t e r i s t i c type o f 'viQ izo m e ' and as such, it can again
move one l e v e l up. This taxonomy i s t h e r e f o r e l e f t w i t h only t h re e
p r o p e r l y included hyponyms o f the hypothesized basic l e v e l term, all
of which are marked for high or low r e g i s t e r . One item 'n a vayo '
( founder , be shipwrecked) may be considered as not p r o p e r l y included
in 'v iQ iz o m e ', since u n l i k e t he l a t t e r verb i t does not n e c e s s a r i l y

219 -
imply ' s i n k i n g ' - in the case o f people on board a ship - but r a t h e r
'be s h i pw re c ke d ' .

T r a n s i t i v e verb taxonomies i nclude even fewer taxa than i n t r a n s i t i v e

ones as can be a t t e s t e d in L i s t V I . The best example o f a t r a n s i t i v e
verb taxonomy i s o f f e r e d by ' p /y en o^ ' ( t ak e to) which appears as a
t wi n inclusive category along wi th ' f e r n o 1 (bring) for the reasons
al re ady mentioned in connection w it h 'p i y e n o j 1 (go) and 'erxom e1
(come).

ferno / piye no2

br ing / take to

metakomizo2 m etafero. metav^vazo axSoforo^j kuvalao

/
move transfer, transmit, c ar r y c a rr y
(furniture) t ra ns por transfer baggage

p e rife ro proskom izo petao^ m etatopizo

c ar r y bring/take gi ve a d i s p la c e
around (a document lift
to an o f f i c i a l )

Fig. 6

Three terms are marked for high register and restricted use:
'a x Q o fo ro ' (carry baggage), 'p r o s k o m i z o 1 (bring, take to) and
'm e taviva zo 1 (transfer). On the o t he r hand, ' k u v a l a o 1 (carry) and
'p e r if e r o ' ( c a r r y around), which are l i n g u i s t i c a l l y unmarked, are not
p ro p e r l y included in the hypothesized hi gher t wi n c at e go r y, because
u n l i k e the cover terms, they need not imply t h a t a d e s t i n a t i o n i s to
be reached. A third factor i s t h e r e f o r e brought i n t o p l a y (besides
linguistic markedness and r e l a t i v e salience of attributes), namely
proper ( o r not) class i n c l u s i o n .

- 220 -
The l a s t taxonomy o f L i s t VI which w i l l be discussed here i s headed
by ' aneveno' (ascend) and i t s near-synonym 'a n e rx o m e '. The l a t t e r
verb is restricted to high register and is in this r espe ct more
l i n g u i s t i c a l l y marked than the former one.

a n e v en o / a n e r x o me

ascend / go upward

ska rfa lo n o ananxome$j a n ifo riz o anabiome^j (an)ipsonome

cl imb cl imb go u p h i l l emerge, rise
( u s u a l l y out o f
wat er )

(sikonom e) ip e n p s o n o m e $ j
rise, stand up r i s e over s . t h

sikonome

rise, stand up

(an)eyirom e$i anasikonome anaka&ome (an)ipsonome

rise, arise s i t up, r a i s e s i t up rise
/lift oneself s l i g h t l y

apoyionome petayom e/petjem e2 iperipsonom e^j

take o f f jump up r i s e over s . t h .

Maui

Only three items are genuine hyponyms of 'a n e v e n o ', namely

'a n arixo m e' (climb), 's k a rfa lo n o ' (cl imb e) and 'a n ifo riz o ' (go up a
slope). The remaining terms, are instances o f p a r t i a l ov er lap r a t h e r
than proper class inclusion. Notice that 'sikonom e' (rise) and
' (an)ipsonom e' (rise), for i nst an ce , can only be replaced by

- 221 -
'a n eveno' (go up) under s p e c i f i c circumstances, e.g . if the moving
o bj e c t is 'smoke' or 'd u s t'. It seems therefore preferable to
consider 'sikonom e' (rise) as independent o f the 'a n e v e n o ' (ascend)
taxonomy and p o s i t it as a basic l e v e l category its elf; it has a l l
the characteristics we a ss oc iat e here w it h lin g u istically unmarked
verbs and i ncludes hyponyms.

Genuine hyponyms o f 'siko no m e ' (rise) are actually those

corresponding to one o f its uses, namely 'ris in g from a s i t t i n g or
lying position', and not its 'change-of-locatio n' use. Partial
overl ap r a t h e r than class i n c lu s i on is r e s po ns i bl e for
substitutability in t h i s s et . This seems t o be very common in most
ot he r castes t e n t a t i v e l y presented as taxonomies in L i s t V I .

To gi ve an o v e r a l l p i c t u r e o f t he remaining contents o f L i s t V I , it
is i mportant to notice that t hey fa ll under t hr ee main c a t e g o r i e s .
One o f them i nvolves hyponyms which are properly included but of
particularly restricted use. In most cases e i t h e r their number is
very smal l , or t hey are so uncommon, t h a t no d e f i n i t e judgments can
be made on them and their exact relation to t he hypothesized
s up er o rd i na t e. In t h i s cat egory f a l l taxonomies 16, 17, 19, 20, 21,
22, 24, 32, 33, 35, 36, 37 and 38.

Another category i ncludes cases of partial ov er l ap rather than

genuine inclusion. Such cases are; 9 (involving onl y one genuine
hyponym), 10 and 23.

A t h i r d category i nv ol ve s sets o f verbs which are headed by a covert

cat egor y, no l e x i c a l i z e d taxon. These sets are c a l l e d here ' n a t u r a l
classes'. Such cases are 7, 12, 13, 18, 25-31 and 34. For some of
these sets a lexicalized hi gher level category is also provided
parallel t o th'e s p e c i f i c a t i o n o f the content o f the c ove rt category
(e.g. 12INTR ' s p e v S o ' [move r a p i d l y ] ) . It is, however, assumed t h a t
such c a t eg or i e s cannot be considered basic l e v e l terms because they
are l i n g u i s t i c a l l y marked.

- 222 -
On the basis of these d a ta , it should be c l e a r by now why it is
claimed that there are only few cases wi th a sufficient number o f
genuine and unmarked hyponyms o f i n c l u s i v e c at e g o r i e s in the area of
MGMVs. Pulman ( 19 83:110) makes a sim ilar observ at ion for English
verbs in g e n e r a l , stating that i t is '’d i f f i c u l t to f i n d enough basic
level verbs w i t h a sufficient number o f hyponyms" i n order to test
whether the pr otot ype e f f e c t obtai ns also f o r verbs.

In this s ec t ion I have concentrated on establishin g p a r t i a l (two

levels deep) taxonomies in the area of MGMVs. The main poi nts
concern the s ta t us of terms pos it ed as ge ner ic and subordi nat e l e v e l
and their relations to one another. It is suggested that the
r e l e v a n t t o o l s f o r accomplishing t h i s ta sk are the noti on o f r e l a t i v e
l i n g u i s t i c unmarkedness and the degree o f class i n c l u s i o n .

The t h e o r e t i c a l di scussion in 4 . 1 . 2 and the d e s c r i p t i o n accompanying

it in 4.1.3, provi de a basis for testing the hypothesis that the
protot ype effect holds for verbs (besides nouns) and suggest what
factors may be r e spons ibl e in this area of investigation for the
formati on of p r ot ot yp e s. Rosch and her colleagues c lai m that the
main f a c t o r i n the case of nouns i s family resemblance (e.g. Rosch
and Mervis 1 97 5 : 5 9 9 ) . Arguments have been o f f e r e d here (1.3.2 and
4 .1) in support o f Pulman's ( 19 83:122) position that the prototype
effect cannot be attributed to family resemblance in the case of
verbs. I t should be obvious from the preceding di scus si on, t h a t even
in the case o f what seem t o be the most promising taxonomies in t h i s
field , we are in obscure territory; (a) we are focusing on
d i f f e r e n c e s between low l e v e l categories (Rosch's subordi nat e l e v e l ,
BBR's specific level) and (b) accurate listings of attributes are
impossible t o compi le. As a l re ady pointed o u t, though, f a c t o r s other
than family resemblance may be s u b s t i t u t e d for it: fam iliarity and
relative linguistic unmarkedness have a lr e ad y been discussed. The
most promising factor seems to be the re la tiv e s al i e n c e of
attributes. The overall issue of which minor properties operate
within the field under investigation and how they relate to one
another need t o be considered in d e t a i l .

- 223 -
4.2 Minor p r o p e r t i e s o f Modern Greek motion verbs

The f a c t t h a t English MVs have been c at egor i ze d in many d i f f e r e n t and

equally plausible ways w i t h i n the framework o f var ious 'checklist'
theories i s i t s e l f an i n d i c a t i o n t h a t the semantic f i e l d in question
i s undoubtedly s t r u c t u r e d but t h a t t h e r e are many angles from which
its structuring can be approached. With t h i s obs er vati on in mind,
the task of identifying common components among MVs in MG (and
probably in any Indo-European language) and b ri n g i n g out their
structural relations is a re latively straightforward m a t t e r . ' The
p itfa ll to avoid i s s t a r t i n g the a na ly s is and basing the d e s c r i p t i o n
on c ont ra st s and opposi ti ons between neighbouring words, which i s the
standard s t r u c t u r a l i s t p r a c t i c e .

A general c l a s s i f i c a t i o n o f MGMVs is presented in L i s t s V I I and V I I I

which does i n v ol v e some s t r u c t u r a l i s t r e l a t i o n s , the exact status of
which can be only discussed after an a n al y s i s of the relevant
p r o p e r t i e s i n v o lv e d. This c l a s s i f i c a t i o n i s only one o f a number of
possi bl e categorizations. Alternatives are discussed at many
d i f f e r e n t poi nts in the course o f the present i n v e s t i g a t i o n .

The main categories identified are the following: 'causatives' of

motion are considered a separate class from 'non-causatives' of
motion. Both cl asses include 'change-of-locatio n' (CL) and 'change-
of-position' (CP) verbs, as wel l as 'change of orientation' ( or
r o t a r y motion) ones. 'General' motion is d i s t i n g u i s h e d from motion
executed in a particular ' manner'. 'D irectional' motion is a
c h a r a c t e r i s t i c o f most CL verbs. 'V e rtica l' motion i s juxtaposed to
motion wi th 'indeterm in ate' direction. Verbs d e sc r i b i n g path/
'passage' and verbs i n v o l v i n g 'dependent motion' appear in groups o f
t h e i r own. A separate category i ncludes verbs marked f o r 'absence of
destination' / 'random w a l k ' . P r op er t ie s i n vo l v i n g the 'medium' /
'environment' i n which the motion is executed, ' i m pe tus ' and ' t y p e of
object' moving w ill be discussed separately. Similar 'minor
properties' are e x h i b i t e d in both 'causatives' and 'non-causatives'
o f motion.

- 224 -
The discussion which f ol l ow s concentrates on non-causati ve CL verbs,
which are by f a r the most numerous. Sim ilarities and d i f f e r e n c e s
between t r a d i t i o n a l approaches and Prototype theory are poi nted out
in the course of this a n al y s i s of MGMVs, such as gr adat io n and
central vs marginal i ns tance s. The term 'minor p r o p e r t i e s ' i mpl ies
here 'specific to the semantic field under investigation'. Their
relationships to one another, relative h i e r a r c h y and d i f f e r e n c e s in
s ta t us cannot be discussed before a p r i o r a na ly s is o f the semantic
content and r e l a t i o n s between a number of verbs understood here as
involving the properties in questi on. The terms 'featu res',
'components', 'p roperties', 'attributes' are used i n t e rcha nge abl y.

Since most o f the di scussion w i l l r evol ve around CL verbs, it seems

in order t o r e f e r t o CP verbs f i r s t which w i l l not be given ext ensi ve
treatment. As can be attested in List V II, three groups are
id e n tifie d within this area. One o f them i ncludes verbs de sc r ib in g
'change o f pos tur e' or 'change o f p o i n t - o f - s u p p o r t ' , e.g. '/ra0 o/ne^'
(s it), ' y er no^ ' (lean), 'y o n a tiz o j' (kneel). Most of these verbs
refer to human body motion. A number of them are also used as
statives d esc ri bi ng the resulting position/posture, e.g. 'kadom ej'
( s i t / b e s ea t e d ) , 'k s a p lo n o j' ( l i e down), e t c .

Whether one of these uses i s more primary than the ot he r cannot be

decided on ' l i n g u i s t i c ' grounds. M i l l e r and Johnson-Laird ( 19 76:549-
50) suggest that since there may be no gross movement in common
between standing from a s i t t i n g posture and standing from a crouching
one, the a ct ion component i s r e l a t i v e l y uns pe ci fi ed and the r e s u l t i n g
posture is the main global concept. On such grounds it could be
claimed that the primary use is the stative one. Bodi ly movement
'change o f p o i n t - o f - s u p p o r t ' verbs appear in L i s t V I I t og e t h e r wi th
verbs sharing t h i s p r o pe r t y w i t h them though these latter ones are
not necessarily human body 'change of pos tur e' verbs, such as
'krem jem e' (be hung), and '\e rn o ' (lean). Notice that if we are
looking for the most constant rather than the most characteristic
pr ope rt y o f these verbs, i.e. a necessary and s u f f i c i e n t condi ti on
r a t h e r than a p r o t o t y p i c a l c o nd i ti o n o f a p p l i c a t i o n , the term 'change
of point-of-support' is more a p p r op r ia te than 'change o f posture'.
Imagine a person's body having a lr e ad y acquired a s i t t i n g posture but

- 225 -
having no seat y e t . This s i t u a t i o n cannot be c a l l e d ' s i t t i n g ' , since
the support i s s till provided by the ground and one's f e e t as f o r a
standing position. It does not require a l engthy e xpl ana ti on to
demonstrate that the necessary c o ndi ti on involves the po i nt of
support. Whether t h i s condition, which covers a l l pos si b le cases,
has any psychol ogi cal validity or not is a different matter. This
issue i s d i r e c t l y l i n k e d t o the d i f f e r e n c e between 'most g e n e r a l ' and
prototypical understanding o f these verbs and w i l l be taken up again
in the next c h ap t er .

A second group o f CP verbs i nvolves also ' p a r t i a l ' motion which does
not , however, result in a different posture of the moving object.
Such verbs are 'k ilje m e ' (wallow), 'kunjem e' (move, s tir), 's a le v o '

( s t i r , move s l i g h t l y ) , and the l i k e , e.g.:

(1) to yurunaki k ilje te s ti lasp i

the p i g l e t wallows in t he mud.

Some o f these verbs cannot appear in 'punctual occurrences' and are

rather odd in event predications in general, others are much less
'processual' . Compare, for i n st an ce , 'k ilje m e ' (wallow) to
'tra d a zo m e ' (jerk, shake).

(2 ) *to yurunaki k ilis tik e s ti lasp i ja mja s t i y m i

the p i g l e t wallowed in the mud f o r a moment

( 3) to s p iti tra d a x tik e ja mja s t i y m i o lo k liro

the house shook f o r a minute 'whol e'
the whole house shook f o r a minute

The d i f f e r e n c e between the second and the t h i r d group i s obvious: the

latter one, involving 're g u lar/re p ea te d ' partial motion, includes
fa irly 'processual' verbs, e.g . 'tre m o ' (trem ble), 'ta la n d e v o m e '

(o scillate), 'p a l o m e ' (vibrate). In Chapter 2 it is shown that

'tre m o ' ( t re m b l e ) p r e d i c a t i o n s cannot be construed ast y p i c a l events.
Expectedly t h e r e are poi nts o f ov er lap betweenthese two l a s t groups.

- 226 -
Consider f o r instance 'a n a p ib a o ' (jump up/jump up and down) which can
i n vol ve e i t h e r a momentary movement or a s e r i e s o f 'jumps'. In the
l a t t e r case ' a n a p i S a o ' i s more l i k e l y to be understood as a CL r a t h e r
than a CP verb, e . g . :

(4a) mol i s akuse ton p ir o v o lis m o a na pibise s tin k a re k la tu

when he heard the shot he jumped up on his c h a i r

(4b) o p r o p o n i t i s e v a l e t u s m a d i t e s na a n a p i b u n
the coach made the p u p i l s jump up and down.

In s h or t , the borders between CL and CP verbs are also fuzzy. In

List VII which presents a classification based on my personal
intuitions in cases o f doubt as t o t he t y p i c a l understanding o f an
item, 'a n a p iS a o ' appears under CP (and as uns p ec if ie d concerning
regularity/repetition of m ot io n) . In the following chapt er ,
di scussing p rototypicality judgments of s u bj e c t s , a more o b j e c t i v e
image o f t h i s verb can be o f f e r e d .

4.2.1 'Change-of-location' and ' d i r e c t i o n a l i t y '

'Change-of-location' (CL) or ' t r a n s l a t i o n ' o f an o b j e c t is considered

in M i l l e r and Johnson-Laird (1976 :5 33) as c o n s t i t u t i n g the "nucleus
o f the semantic domain" o f MVs. The f u l l c o n st r u c t i o n they propose
can be s i m p l i f i e d , f o r our present purposes, and presented as (FROM
(TO (TRAVEL))) (x ,iv,v). TRAVEL i s used as t he most general relevant
predicate, i.e. as expressing CL in the s impl es t way p os si bl e (see
a lso M iller 1 97 2 ). Prepositional phrases appear as predicate
modifiers, so w and v r epr es ent the in itial and final locations
respectively of t he moving object (x). The main c at e gor i es
identified within this sub-area of MGMVs appear in List V III in
diagrammatic form.

T h e o r e t i c a l l y any change o f l o c a t i o n can be understood as a d e v i a t i o n

from the axes set on a plane or in space (Ikegami 1969:112). In the
most general terms possible, one would expect language to express

- 227 -
motion along a v e r t i c a l axi s (i.e . in r e l a t i o n to g r a v i t y ) , along a
horizontal axis (e.g. forward - backward), around an axi s (rotary
moti on) . N ot i c e , however, th a t unlike v e r t ic a l direction, 'forward'
and 'backward' in f a c t express d i r e c t i o n r e l a t i v e t o notions such as
an o b j e c t ' s n a tu r a l f r o n t or h a b i t ua l d i r e c t i o n of motion. Similarly
'h ere' and 'there' i nv ol ve direction relative to speaker and
addressee's l o c a t i o n , i.e. motion r e l a t e d to d e i c t i c elements o f the
language also apparent in the use o f d e i c t i c verbs such as ' p i y e n o j '
(go) and 'erxom e' (come). If the po i nt of o rig in and t he po i nt of
d e s t i n a t i o n are marked f o r being on d i f f e r e n t l e v e l s , verbs are used
which are marked for v erticality, e.g . ' aneveno' (ascend) or
'v u lja z o ' (sink). There are no verbs, however, which are
cor respondi ngl y restricted to horizontal direction. If a
classification of MVs is made on the basis of their most general
understanding, a distinction can be drawn between verbs marked f o r
v e r t i c a l i t y and verbs which are i n de te rm ina te i n terms o f d i r e c t i o n a l
orientation (i.e . unmarked for vertic a lity) rather than between
vertical and horizontal direction specifying verbs. No t i c e , for
i ns tan ce , that 'p r o x o r o ' (advance), 'b e n o ' (enter), ' ip o x o ro '
(withdraw) are not restricted to a horizontal axis in the same way
th at 'sk a rfa lo n o ' ( cl i mb ) is restricted to a vertical one (unlike
English ' c l i m b ' of cour se ), e . g . :

(5) ; ipo xorise s t in k o r i f i tu v unu

s/he r e t r e a t e d t o the top o f the mountain

may be p r a g m a t i c a l l y odd but is not unacceptable, and

(6 ) proxoruse o lo ke p j o vaQja s t o p i y a b i

s/he advanced more and more deeply i n t o the wel l

is a p e r f e c t l y acceptabl e sentence. For t h i s reason verbs marked f o r

presence o f v e r t i c a l i t y are c l a s s i f i e d s e p a r a t e l y in L i s t V I I I and no
corresponding 'h o rizo n ta lity ' box is envisaged. If prototypical
c ondi ti ons of application are taken into account, though, it is
possi bl e t o i d e n t i f y typical uses o f ' ip o xo ro ' (withdraw) which may
be l i n k e d w it h e i t h e r a h o r i z o n t a l or a 'downward' d i r e c t i o n but not

- 228 -
wi th an 'upward' one. Such di scr epanci es between a general and a
prototypical understanding in terms of directionality are most
e vi de nt in the case o f some 'manner' s p e c i f y i n g MVs and w i l l be taken
up and discussed in d e t a i l a t a l a t e r p oi n t in t h i s c h ap t e r .

Directional specifications ot he r than vertical ones do not appear

separately in List V III. Their importance is not at i ssue. Most
analyses of English CL verbs provi de or presuppose an ext ensi ve
t re at ment of directional a d ve r b i al s and l o c a t i v e p r e p o s i t i o n s which
are t r a d i t i o n a l l y understood as semantic components o f the verbs in
question (e.g. Gruber 1965, 1976, Ikegami 1969, M iller 1972, M iller
and Johnson-Laird 1976).^ The absence of sim ilar s t udi es for the
corresponding MG d i r e c t i o n a l / l o c a t i v e Advs renders impossible the
t a sk o f anal ysing such semantic components o f CL MG Vs in depth and
b ri ngi ng out t h e i r p r o t o t y p i c a l characteristics. T h e i r combinational
possibilities, as wel l as t h e i r relative salience in comparison wi th
ot he r kinds o f semantic m a t e r i a l (e.g. 'manner', 'c a u s a tiv ity ') w ill
be discussed in 4.2.4, 4.4 and e s p e c i a l l y in Chapter 5. At this
point, examples o f MG CL verbs w i l l be provided which can be s a f e l y
matched wi th directional/locative Advs as analysed in Ikegami
(1969:112-31). F ur the r examples can be found in List VI; the
addresses of the r e l e v a n t groups o f verbs appear under each one of
the Advs presented in what f ol l ow s immediately:

D i r e c t i o n a l A d v er bi al s Examples o f MGMVs

upward / up 'aneveno' (ascend)

6,7,8 ' a p o y i o n o m e ' ( t ak e o f f )
' s k a r f a l o n o ' (c li mb)

downward / down ' k a t e v e n o ' (descend)

9,10,11,19,33 'k a tra k ila o ' (roll down)
' v u l j a z o ' (sink)

- 229 -
 Directional Adverbials Examples of MGMVs

forward / onward 'p ro x o ro ' (advance, proceed)

35 'prooQume' (advance)
'p ro e la v n o ' (push forward)

backward / back ' o p i s Q o x o r o ' (move backward)

29,30,36 'opisQ oftrom o' (retreat)
'e p is tre fo ' (return)

inward / in 'b e no ' (enter)

19,21,33 'is v a lo ' (i nvade)
'v u ta o ' (dive)

outward / out 'vyeno' (move out o f)

22, 37 'p ro v a lo ' (appear out of)
'ekserxom e' (come out o f)

towards / to 'p lis ia z o ' (approach)

'proserxom e' (come t o )
' episkeptom e' (visit)
' fQ ano' (arrive, reach)

away from / from 'a p e rxo m e ' (go away)

1a p o x o r o ' (withdraw)
k s e k in a o ' (s ta rt off)

Besides these d i r e c t i o n a l adverbials, a number o f the aforementioned

analyses of English MVs, i ncl ude also 'through', 'by' and 'a c ro ss '
which r e f e r t o i n t e r m e d i a t e l o c a t i o n s o f the j our ne y, 'around' which
r e f e r s to 'change o f o r i e n t a t i o n ' , 'w ith ', 'after' and ' b e f o r e ' which
refer to motion of an o b j e c t relative to the location(s) of some
ot he r o b j e c t . These w i l l be considered and e x e m p l i f i e d separately;
they are understood here as being o f a d i f f e r e n t s ta t us than the ones
a lr e ad y mentioned. The r e l e v a n t sections in L i s t V I I I are l a b e l l e d

- 230 -
'Path', 'Dependent Motion', 'Random Walk' and 'Change of
O rien tatio n'.

4.2.2 'Path' and 'dependent motion'

The term ' p a t h ' i s used here in a r e s t r i c t e d sense u n l i k e i t s use in

M iller and Johnson-Laird (1976) or Fillmore (1978) where 'pa th of
motion' i ncludes e ve r y t h i n g which i s here understood as d i r e c t i o n a l
change-o f-location. It i s juxtaposed to ' s ou r c e ' and ' g o a l ' i n the
sense t h a t i t concentr ates on i n t e r m e d i a t e l o c a t i o n s o f the j our ne y.

Verbs belonging to t h i s category appear under 24 o f L i s t V I . Since

the verbs in questi on de scr ibe in essence motion relative to some
specific location usually taking up the position of their direct
object, they can be r e l a t e d t o a separate class o f CL non-causati ve
transitives which do not descr ibe 'path' (or 'passage') but are
sim ilar to them in d e s c r i bi n g motion relative to some location
occupying the position of their direct object. These are:
'p e r i k i k l o n o ' (en circle), 'triy in 'z o 2' ( s u r r o un d) , 'parakam pto'
(deviate, pass around), e t c . , e.g .:

( 7) o s t r a t o s p e r i k i k l o s e tin p o li

the army e n c i r c l e d/ s u r r ou nd e d the town.

Such verbs c o n s t i t u t e probably a l i n k between c e n t r a l cases o f ' p a t h '

(e.g. 'p e rn a o ' (pass), 'b ia s x iz o ' (traverse)) and marginal cases o f
the 'change o f o r i e n t a t i o n ' group ( t o be discussed a t a l a t e r p o i n t ) .

In List V I I I verbs d e s cr i b i n g motion r e l a t i v e to some o t h e r moving

e n t i t y r a t h e r than a f i x e d l o c a t i o n appear under 'dependent m o t i o n ' .
These are also understood as l i n k e d to the group discussed in the
preceding paragraph, in the sense t h a t 're la tio n a l' characteristics
are again central to their meaning. The English c o un te r pa rt s of
these verbs are commonly analysed as incorporating 'directional
a dve rbi al components' of var ious forms, e.g. 'before', 'a fte r',
' to g e th e r'/'w ith '.

- 231 -
A l l o t t i n g MG verbs t o such classes i s as simple a t a s k as i t is f o r
t he e q u i v a l e n t English verbs and examples are provided here f o r the
sake of completeness rather than anything else. So verbs like:
' akoludo' (follow ), 'kiniyao' (run a f t e r , hunt), 'katabioko' (chase)
are trad itio n ally understood as i n v ol v i ng 'a fte r'. Verbs like
'obiyo' (lead), 'proiyume' (go in front of) 'proporevome' (walk in
front of), 'p r o t r e x o ' (run in front o f ) may be said to i nvolve
'be fore '. All these verbs are transitive non- c au s at iv e s. Some o f
them are marked f o r 'manner' as well as ' r e l a t i v e / d e p e n d e n t m ot i on ' ,
e.g . 'p r o p o r e v o m e ' (walk (a long d i st an c e) in f r o n t of), 'p r o t r e x o '
(run in f r o n t o f ) , ' p r o v a b i z o ' (walk in f r o n t o f ) .

As a l re ady noted in Chapter 3, neither 'kiniyao' (run after, hunt)

nor 'katabioko' (chase) require that th eir direct object be moving,
in c o n t r a s t t o 'akoluBo' (follow ). There i s no doubt, however, t h a t
their prototypical uses i n vol ve such a c o n d i t i o n . Notice, simply,
that a classification based on the 'most g e n e r a l ' meaning of verbs
would have no j u s t i f i c a t i o n for a l l o t t i n g them to this particular
group. The s i t u a t i o n i s d i f f e r e n t in the case o f 'obiyo' ( l e a d ) the
prototypical i nstances o f which do not i nvolve a c o n d i t i o n t h a t the
subj ec t should be ' i n f r o n t of/ ahead o f ' the o b j e c t , e.g.:

( 8 ) t o n o b i y i s a n s t o a s t i n o m i k o t mi ma
they l ed him to the p o l i c e s t a t i o n .

In terms o f t he prototypical image conveyed by (8) there is little

doubt t h a t the v i c t i m and his guards are side by s i d e. The relevance
of t h i s obs er vati on l i e s simply in the f a c t t h a t n e i t h e r the verbs o f
the 'a fte r' group nor 'obiyo' (lead) of the 'in front of' group
c oncentr ate on physical motion as such. They i n v ol v e condi ti ons of
a p p l i c a t i o n which focus on o t he r c h a r a c t e r i s t i c s such as 'purpose',
'in te n tio n ' and the l i k e . T h e re f or e, 'incorporation' of d irec tio n a l
a d v e r b i a l s does not seem t o be the c en t r a l issue in connection w it h
quite a number o f these verbs and e s p e c i a l l y the commonest ones. The
group of verbs trad itio n ally analysed as involving 'together' or
'with' c o n s t i t u t e perhaps a c l e a r e r example o f the i na ppr opr ia te ne ss
of the noti on of 'incorporatio n of directional Advs'. It is

- 232 -
quest ionabl e whether 'w ith' and ' t o g e t h e r ' can be classified as
directional adverbials/prepositions (such as 'up', 'o u t', 'before',
'acro ss'). T y pi ca l examples of this group ar e: 'sinobevo'
(accompany), ' simporevome' (walk (a long d i s t an c e) t o g e t h e r w i t h ) and
its near-synonym 'sinobiporo' and 'sintaksibevo' (travel t oge the r
w ith),

Some verbs are also marked f o r ' manner', e.g. 'simporevome' (walk
together w i t h ) . The commonest verb o f the group, namely 'sinobevo'
(accompany), has no additional specifications, either in terms of
type o f motion, type o f moving o b j e c t , medium/environment in which
motion takes place or d i r e c t i o n a l i t y (in terms o f a xe s) . Moreover,
it i s a p p l i c a b l e t o any kind o f a c t i v i t y / a c t i o n , e .g .:

(9) o m i x a l i s trayubai ke i anita t o n s i n o b e v i me t i n kiBara

Michael sings and Anni ta accompanies him wi th the g u i t a r .

It can be argued, of course, that its prototypical i nstances do

i n vo lv e physical CL ( o f some d u r a t i o n ) which i s c a r r i e d out so t h a t
the 'principal traveller' (taking up the position of the direct
object) would not be l e f t alone. The most general understanding of
t h i s verb i s not however l i n k e d t o ( p h y s i c a l ) CL.

The pr ope rty o f 'accompaniment' can be understood i n a more general

way and used to d i f f e r e n t i a t e between a l l 'dependent motion' verbs
and those t y p i c a l l y involving 'i mpe tus ' or ' p r o p e l l e n t motion' such
as 'ektoksevome' (be launched), 'petayome' ( s pr i ng up) and i t s n e ar ­
synonym ' t i n a z o m e ' , ' e k t o k s e v o m e ' (be l aunched). Gruber (1965, 1976),
for i n st an ce , classifies all verbs involving 'before', 'a fte r' and
'with' t o g e t h e r as i n c o r p o r a t i n g 'accompaniment', a l a b e l which seems
more suitable to the very last sub-group mentioned her e, i.e. MG
'sin -' compounds. The r elevance of such a broad distinction, in
terms o f presence or absence o f 'accompaniment', is more apparent in
t he case o f c aus at ive s o f motion, where a l a r g e cat egory o f verbs can
be characterized as lacking this p r o pe r ty . Such a cat egory would
i n v ol v e not only verbs marked for ' i mpe tus ' (e.g. the causati ve
c ount er par ts of the above mentioned set) but also others such as:

- 233 -
'ste ln o ' (send), or 'a poyiono' (cause a plane to take off).
'Accompaniment' could then serve in i d e n t i f y i n g c o n t r a s t p a i r s , e.g.
'piyeno2 ' ( t ak e t o ) vs ' s t e l n o ' (s end). In the area o f caus at ive s o f
motion, presence of 'accompaniment' is probably characteristic of
most CL verbs, e . g . groups 15, 17, 28, 29, 31, 34, 37 o f L i s t V I . It
is a necessary pr op e rt y o f the d e i c t i c pair ' p / y en o^ ' (take to) and
'ferno' ( b r i n g ) and some o f t h e i r hyponyms, e . g . 'proskomizo' (bring/
take (a document to a civil service o ffic ia l)), 'kuvalao' (carry
s.th. (heavy)), 'p e rife ro ' (carry around), etc. It is probably
const ruabl e as one o f the p r o t o t y p i c a l c ondi ti ons o f a p p l i c a t i o n of
causati ves like 'vazo' (put), 'vyazo' (take out) and a number of
t h e i r hyponyms, e . g . 'xono' (stick into). Groups 13TR and 20 in L i s t
VI are c h a r a c t e r i z e d by i t s absence.

The r e l e v a n t distinctions between c au s a t i v e s , in this respect, are

considered in d e t a i l in Chapter 3, where i t i s argued t h a t the notion
of 'control' (which is understood here as subsuming the notion of
'accompaniment') is a b e t t e r tool in b ri n gi n g out such d i f f e r e n c e s ,
in connection wi th c a u s a t i v i t y and a g e n t i v i t y . At t h i s point it is
only worth n o t i c i n g t h a t 'accompaniment' as a ' mi nor ' p r o pe r ty seems
to have a d i f f e r e n t status in the c au sa ti ve group than in the non­
caus at ive one, as i t is applicable to a f a r g r e a t e r number o f verbs.
Its s al i e n c e relative to o t he r ' m in or ' properties can be discussed
again in connection w i t h some t e s t r e s u l t s i n the f o l l o w i n g c hapt er .

4.2.3 'Change o f o r i e n t a t i o n '

The term 'change o f o r i e n t a t i o n ' i s a p p l i c a b l e t o two types o f verbs.

The f i r s t type are u s u a l l y r e f e r r e d to as ' r o t a r y motion' ver bs. The
commonest members o f the group, i.e. 's triv o j' and 'y iriz o j' (both
implying 'tu rn ') are not n e c e s s a r i l y l i n k e d wi th (complete) r o t a t i o n
as such, but r a t h e r w i t h change o f d i r e c t i o n , e.g.:

(10) sti yonia t u b r o mu e s t r i p s e b e k s i a

a t the c orner o f the s t r e e t s/he turned r i g h t .

- 234 -
The next commonest one' s t r i f o y i r i z o ' (twist and turn) appears
t y p i c a l l y in an environment o f the type:

(11) b e n b o r u s e na k i m i Q i ke s t r i f o y i r i z e sto k re v a ti

s/he could not sleep and t w i s t e d - a n d - t u r n e d in bed.

None of these characteristic uses is central to the noti on of

rotation around an e x i s t i n g or not iona l axis although 's t r i f o y i r i z o '
(twist and t u r n ) in (1 1) is closer to (irregular and incomplete)
rotation than 'strivo j' (turn) in ( 1 0 ) . This is perhaps better
presented in l ess common verbs o f the group, e.g . ' p e r i s t r e f o me '
(revolve, twine), 'peribinume' (whirl, swirl, eddy), 'strovilizom e'

(w hirl), 'elisome' (snake, w i n d) , 'p e ri elisome' (coil), 'kulurjazome'

(curl up, w h e el ), e t c .

Most o f the verbs o f r o t a r y motion do not d i s t i n g u i s h c l e a r l y between

change-of-location and c h a n g e - o f - p o s i t i o n . Consider, for i ns tance,
' anapoboyirizo' and its near-synonym 'tubaro' (overturn). So they
d i f f e r from the classes a lr e a d y discussed in more than one r e sp e ct .

Another group of MGMVs which can be also understood as i n v o lv i n g

'change o f o r i e n t a t i o n ' , but no r o t a r y motion, i ncludes verbs which
are indisputably CL ver bs, and characteristically ' p r oc e ss u al '
(d u r a t i o n is a definite characteristic o f the whole gr oup) . Change
of direction is in their case 'random' unlike the cat egory just
discussed, and irregular, on t he whole. The ne ar es t notion in
physics seems t o be 'random w al k ' which i mpl ie s t h a t the past h i s t o r y
(direction) of the motion is not r e l e v a n t to its continuation, i.e.
each next step depends only on the location previously held and
cannot be p re d ic te d on the basis o f what has preceded i t . The term
'random wal k' is used in List V III to d i s t i n g u i s h these verbs from
those involving rotary motion which appear under 'change of
orien tation'. T h e i r most s a l i e n t c h a r a c t e r i s t i c seems to be 'absence
of final d e s t i n a t i o n ' or 'a i ml es s change o f l o c a t i o n ' . The group has
an unusually g r e at number o f items, the commonest ones being:
'triy iriz o ' (roam around), 'periplanjeme' (wander) and 'periferom e'

(r ov e, roam around) which do not s p e c i f y 'manner', 'type o f object'

- 235 -
moving, ' i nstr ument/ means' o f motion. A subset i nv ol ve s hyponyms o f
'perpatao' (walk), e.g . 'sulatsaro' (stro ll) and its near-synonyms
'peribjavazo' , 'se rja nizo' , ' v o l t a r o ', etc. One verb 'armenizo'
(sail about) requires a specification of 'th e sea' as a 'medium/
environment'.

The main interest of this group lies in that it is extremel y

d ifficu lt to interpret its members in terms of necessary and
sufficient conditions, contrastive features and formal components,
although the semantic s i m i l a r i t y o f i t s members i s obvious. Even the
term ' ai mless change-of-location', which is suggested here as
r e f l e c t i n g t h e i r most s a l i e n t p r o p e r t y , is ra the r unfortunate. It is
c l o s e r t o the f a c t s to suggest t h a t d i f f e r e n t grades o f 'absence o f
purpose' or 're lativ e seriousness of purpose' can be identified
w i t h i n t h i s group.

Verbs like 'p e r i i y u m e ' (tour) and 'p e r i o b e v o ' (tour, travel), like
' taksibevo' (travel) i t s e l f would be a t the top of a 'seri ousness of
purpose' s c al e , in comparison to the remaining verbs of the group.
The verb 'periplanjeme' (wander) could be immediately a f t e r them, in
t h i s r e sp e ct , as i t does not s p e c i f y whether wandering is i n t e n t i o n a l
or acci dental (due to l o s in g one's way r a t h e r than wandering around
f o r amusement).

The verbs mentioned in the f i r s t p a r t o f the di scussion o f t h i s group

(e.g. 'periferome' (roam around)) are a step lower on the scal e and
there are differences between them too, all linked to the
s u b j e c t i v e l y assessed pr op e rt y o f ' seriousness o f p ur pos e' . N ot ic e ,
for i ns tan ce , that while ' kano volta' (have a walk) has a fa irly
p o s i t i v e c o nn ot a ti o n, 'voltaro' (s tro ll), though also e t y m o l o g i c a l l y
related to it, has a r a t h e r n egat i ve connotation and 'kovo voltes'

(walk around) is even worse. A great number of verbs and verbal

expressions which belong to the domain under consideration have a
d i s t i n c t l y ne ga t iv e connotati on and could be t h e r e f o r e understood as
l y i n g at the bottom o f the suggested scale where 'complete absence of
seriousness of purpose' is matched wi th soci al disapproval. Such

- 236 -
verbs are: 'yiroferno' , 'yirnovolao' , ' a l o n i z o ', 'vosko',

'surtukevo' , 'a lite v o 't ' k o p r o s k i l i z o ' , etc.

I t can be e a s i l y argued, o f course, t h a t such c o n si d er a ti o ns are o f a

pu r el y sociolinguistic nature and t h a t no semantic t h e o r y needs to
account f o r them. The p o i n t made here, however, is that presence,
absence and re la tiv e seriousness of purpose are the most salient
properties of this group. They are r esponsi bl e for the semantic
sim ilarity of the items in questi on, they constitute the very
p r o p e r t i e s which d i f f e r e n t i a t e them e s s e n t i a l l y from o t h e r groups o f
MGMVs and they cannot be adequately replaced by concomitant f e a t u r e s
such as change of direction and d ur at ion of the motion. It is
f u r t h e r suggested t h a t w i t h i n the framework o f p rot ot ype t he or y such
properties can be easily accommodated in the form of descriptive
condi ti ons o f a p p l i c a t i o n ( r a t h e r than ' y e s- n o' c riteria ! features).
Such c ondi ti ons have the additional advantage of not requiring
homogeneity, i.e. they can be of different types and can combine
p e r c e p t u a l l y w i t h c u l t u r a l l y important i n fo r m a t i o n .

4.2.4 ' Ma nn er ', 'medium', and ' i n s t r u m e n t a l i t y '

In what has preceded, r e f e re n ce is of t en made to verbs belonging to

vari ous c a t e g o r i e s , such as 'vertical direction' or 'dependent
motion', w h i l e a t the same time s p ec if yi ng 'manner' o f motion, e.g.
'skarfalono' (climb), or ' s i m p o r e v o m e ' (walk (a long d i st a n c e )
to g e t h e r w i t h ) . The d i s t i n c t i o n between ' gener al moti on' verbs and
'manner s p e c i f y i n g ' verbs i s not , however, a w i d el y accepted one. In
fact, analyses o f English MVs use 'manner' as a l a b e l f o r much f i n e r
distinctions than the ones i mpl ied here, e.g . Fillm ore's (1978)
'manner' issued for differentiating between 'strid e' and 'scurry',
w h i le M iller and Johns on- La ir d' s (1976) example of 'manner' is
'travel rapid ly'. In the pr esent analysis, 'manner' is used in a
fa irly broad sense and is meant to cover t h r e e different areas: a
central one i n v o l v i n g vari ous types o f (typically) human motion on
ground, e.g . 'tre xo' (run), 'p e r p a ta o ' (walk), a much wi der area
involving d i f f e r e n t ways o f moving in water or a i r , e.g. 'k o lib a o '
(swim), ' p e t a o j ' (fly ), and a most restricted one i n vo l v i n g

- 237 -
distinctions at a low levelo f i n c lus ive ne ss as exemplified by
hyponyms of verbs belonging to the central area, e.g . 'vim atizo'
(p a ce ) , 'b r a s k e l i z o ' (s trid e ).

The r e l a t i o n s between 'ma nne r' , 'medium' and 'instrum entality' are
presented d i a g r a m m a t i c a l l y in F i g . 8 ( o v e r l e a f ) , where the innermost
circle i ncludes 'general' MVs and the next larger one, verbs
specifying manner in a broad sense (involving instrumentality and
medium). The verbs appearing o ut si de this larger c irc le belong to
the most s p e c i f i c understanding o f 'ma nne r' . Both l a t t e r c a t eg or i e s
are also classified on the basis of the medium/environment within
which the motion takes pl ace .

The reason f o r consi de ri ng such w i d el y d i f f e r e n t kinds o f motion as

belonging t o g e t h e r and as separate from 'ge ne ral motion' verbs is
t h a t they behave in a s i m i l a r manner when combinedw i t h verbs o f the
latter group. Notice, for i n s t an c e , that all the verbs s p ec i f y i n g
'manner ', which are given above as examples, can appear in the
f o l l o w i n g environment i n the form o f Present P a r t i c i p l e s :

(12) piye/efBase (kapu) _________

s/he went t o / a r r i v e a t (some pl ace ) ________

The final position in the above environment can be occupied by:

'trexondas' (running), 'p e r p a t o n d a s ' (w alking), 'busulizondas'
( c r aw l i ng (as of a baby)), 'kolibondas' (swimming), 'petondas'

(fly in g ), etc. The reverse (in terms of grammatical forms) is

i mpossible:

(13) *kolibise mexri to vraxo piyenondas

s/he swam up t o the rock going.

I t has t o be e xpl ai ned a t t h i s p o i n t t h a t 'manner' i s understood here

as r e l a t e d both t o 'medium' and ' i n s t r u m e n t a l i t y ' . The term 'medium'
is o ft e n used to indicate the environment in which motion takes
place, i.e. l a n d, w a t e r , air. 'In stru m entality' i s used to i n d i c a t e

- 238 -
 CO

£ 3
>
co 32 c+ N
m s

d-

CD < (D O

ro.|vi fu

Cfl Tea
c i- O O

£2 (T>
O O O
c + - f i CD

52=
30 O CD

cn fD O C
3 3
c + iD
fD
3
CHANGE OF MEDIUM
03 fD - O

CO v u ta o 1( d iv e )
to 3 O
fD fD
Ul "O f + t l 32
7T ~S fD CO
CD Ol fD -S — I
~5 ED >C - a po
-h O O CD CZ
CD - - r+ 3
O <3-1
.-S o

o <
E fD

C-J. Ol
—i 3 p>
C -J. O 7T
fD 3
O fD 3 fD — ■
s
m
o -s

< C
D
O* _i
c 3 -O

-o
N O

O
CD
the means/cause o f motion, i.e. human body, conveyance and perhaps
gravity. It must be noti ced t h a t MG does not nor mall y l e x i c a l ize
i n s t r u m e n t a l i t y through cover terms such as English 'ride' implying
moving by c ar , f o r i ns t an ce , or ' f l y 7 implying moving by pl ane. Even
'pleo' (s ail) and i t s hyponyms, e.g. 'apopleo' (sail off/away), are
mainly used to desc ri be the motion o f the ship and r a r e l y t h a t o f the
people on board. In short, instrumentality is commonly expressed
through PPs o f the form: ' me t o aeroplano/plio/aftokinito' (by p l a n e /
ship/car). The same type o f PP i s used in j u x t a p o s i t i o n t o the ones
just mentioned to specify 'fe e t' as instr ument , i.e. 'me t a p o b j a '
((lit.) w it h the feet, on foot). S trictly speaking, this last
expression does not d i f f e r e n t i a t e between running and walking but i t
is in fact the latter kind of motion which is basically i mpl ied,
since 'walk' is the unmarked member o f the p a i r , i.e . the commonest
way o f changing l o c a t i o n (on land and) on f o o t .

The interdependence o f 'medium' w it h 'instrum entality' and 'manner'

is also f a i r l y c l e a r , although 'medium' cannot be compl etel y subsumed
under 'manner' in the way t h a t 'instrum entality' can. N ot i c e , for a
start, that the majority of 'body i nvolvement' CL non-causatives
imply ' l a n d ' as a medium ( e . g . 'sernome' (creep), 'ka trakilao ' (roll
down), 's k a rfa lo n o ' (clim b)). Some o f them imply 'loss o f contact
w i t h ground' and may be considered as i n v o l v i n g both land and a i r as
'medium', e . g . 'p i b a o ' (jump).

A number o f verbs exist f o r which 'water' as 'medium' needs to be

specified. Verbs implying motion in/into wat er are, for instance,
'viBizome' (sink), 'vuljazoj' (sink), 'vutaoj' (dive). Verbs
implying motion on/in wat er ar e: 'kolibao' (swim), 'pleo' (sail),
' armenizo' (sail a b o ut ) , etc. There are also examples of verbs
implying motion out of wat er : 'anafiiome' (emerge ( out of water),
break w a t e r ) . Two of these verbs r e q u i r e speci al mention: 'vutaoi'

(dive) and 'kolibao' (swim). Besides motion in w a t e r , among ot he r

things, 'vutaoj' (dive) i mpl ie s ' l o s s o f contact wi th ground' and can
be therefore understood as an in between case, a cross of two
mediums: land and w at er ; it is also typically (although not
necessarily) linked to human body motion. The latte r verb, i.e.

- 240 -
'kolibao' (swim) is probably the only central case o f 'manner' and
'body i n v o l v e m e n t ' , though i t also c h a r a c t e r i z e s the motion o f f i s h .

The remaining verbs could, in p r i n c i p l e , be used f o r d i f f e r e n t types

o f obj ect s moving, and the 'manner' o f motion can, in t h e i r case, be
understood as the result o f the combination of 'water' (as medium)
and ' k i n d o f o b j e c t ' moving, but nothing more s p e c i f i c than t h a t . It
is therefore ar guabl e, that wi th the exception of ' k o l i b a o ' (swim)
and perhaps 'vutaoj' (dive) and i t s near-synonym 'kataS/oroe', the
remaining verbs do not r e a l l y specify 'manner' of motion but other
properties such as 'medium', 'd ire ctio n a lity ', 'purpose', ' t y pe of
o bj e c t m o v i n g ' .

Similar c o n si d er a ti o ns are applicable in t he case of verbs which

require 'a ir' as medium. The c ount e rpa rt of 'kolibao' (swim) is
'p e ta o j' (fly ) in this area, which i mpl ie s 'body involvement' and
specific 'manner' of motion for bi rds but is otherwi se sim ilar to
'pleo' (s a il) in that it is also used f o r o t he r kinds of objects
moving in t he a ir, typically planes, marginally objects in them.
Verbs l i k e ' apoyionome' ( t ak e o f f ) and 'prosyionome' (land) i nvolve
two mediums, land and a i r ( s i m i l a r l y to 'vutaoj' (dive)). The same
appl ie s to a l l the verbs implying ' i mpetus' or ' p r o p e l l e n t m ot ion' ,
e.g . 'ektoksevome' (be l au nched), 'tinazome' (spr ing up), etc. This
last group is a f u r t h e r example of problems of general
classificatio n s, or rather t h e i r subjective character. Verbs
implying ' i mpe tus ' can be classified under 'ma nne r' , 'change of
medium' (if such a category needs to be invented), 'absence of
accompaniment' or 'general-d ire ctio n al' motion. Notice, further,
that 'loss of c ont act w it h ground' is also applicable to 'piSao'

( jump), 'xoropibao' (hop) and the l i k e .

The po i nt o f view o f the c l a s s i f i c a t i o n is necessarily a m at te r of

choi ce, unless a prior a n al y si s of specific items t o be c l a s s i f i e d
r ev ea ls which properties are more salient than o t he r s , e.g.
whether, for i ns t an c e, the pr ope rt y o f 'manner' (clambering motion)
i s more s a l i e n t in 'skarfalono' (c li mb) than the pr op e rt y o f 'upward'
direction or v i c e v er sa . At t h i s point, different possibilities of

- 241 -
classification are presented w i t h a view t o accommodating p r o p e r t i e s
which are here understood as l i n k e d to 'manner' o f motion. 'Medium',
'instrum entality' and ' i mp e tu s ' are regarded as being most closely
related to 'manner' in the sense t h a t t h e i r interaction results in
various types o f locomotion. Reference to 'type of object' moving
has also been made in this connection. It is concei vabl e that a
general classification should be based on ' t y p e of object' moving,
since a lot of 'manner' distinctions depend on it. The most
prominent examples i n vo lv e verbs d e sc r i b i ng the motion of liq uid s/
masses, e.g . 'stazo' (drip), 'xinome' (flow into, be s pilled).
Not ic e, also, that differences in 'speed' exemplified through
subordinate l e v e l items such as 'aryosalevo' ( s tir slo w ly/slightly),
'a n o kila o ' (flow slowly), ' t a x i p l o o 1 (sail rapid ly), etc. should
also be subsumed under 'manner' in the understanding of 'manner'
posi ted a t the beginning o f t h i s secti on as the most r e s t r i c t e d one.
Other examples of high specificity i n vol ve 'length of di st an ce
covered' as exhibited by hyponyms of 'p e r p a t a o ' (wal k) such as
'porevome' (walk a long distance) and combinations of 'speed and
length' pr esent in verbs like 'vraSiporo' (walk a long di stance
slowly).

Following M i l l e r and Johnson- La ir d' s (1976:550) t e rmi nol ogy, we could

consider verbs such as 'p e r p a t a o ' (walk), 'trexo' (run), 'sernome'

(crawl, creep), 'skarfalono' ( cl i mb ) and 'k o li b a o ' (swim) as

referring to "main gl obal locomotory motions". The equivalent
English verbs have re ce ive d e xt e ns ive tr ea tment in many d i f f e r e n t
ways and the e x e r c i s e i s not worth r e p e a t i n g f o r MG verbs. Specific
poi nts o f v ar ious approaches a r e , however, worth di scussing in order
t o show the d i f f e r e n c e s between them and Prototype t he o r y .

A p os si bl e approach t o t h i s p a r t i c u l a r set o f verbs is presented by

Leech (1 96 9: 1 89 ) who e s t a b l i s h e s a " m u l t i p l e taxonomic system" o f the
form:

1 LOCO 2 LOCO 3 LOCO etc.

'walk' run
'ru n' 'crawl'

- 242 -
 The f i r s t two examples correspond to 'go on f o o t ' , the t h i r d one to
'go on all fours'. The terms 1 LOCO, 2 LOCO, etc., stand f o r
different ways o f changing location wi thout involvement of any
ex t er na l instr ument . Such an approach concentrates on the s t r u c t u r a l
relations o f the items in question and seems to be c o n s i s t e n t wi th
the p r i n c i p l e o f the ' maxi mal ly g e n e r a l , minimal s p e c i f i c a t i o n s ' (see
Chapter 1 ) . It is fa irly easy to show t h a t anything more s p e c i f i c
than such incomplete d e f i n i t i o n s risks v io la tin g th is principle. It
can be also argued t h a t even l engthy and comprehensive accounts o f
the verbs in question result in incomplete and unsatisfactory
definitions if they are governed by the i d e a l s of neat formulae and
maximal generalization. Consider f i r s t , as a case in p o i n t , M iller
and Johnson- La ir d' s ( 1 9 7 6 : 5 47 - 53 ) account of 'w alk'. The motion
labelled by 'walk' is analysed in a complex o f lower l e v e l bo di l y
movements, such as liftin g a foot from the ground and moving the
ot he r one in f r o n t of i t w hi le simul taneously moving the body forward
and changing its c e nt r e of gravity, etc. To d i f f e r e n t i a t e between
'run' and 'w alk', M iller and Johnson-Laird (ib id .) introduce a
component ATG which stands for 'always touching gr ound' , refe rs to
'fe e t' and i s p a r t o f the s p e c i f i c a t i o n of the l a t t e r verb but not o f
the former one. ATG i s meant as an i l l u s t r a t i o n of "how manner of
travelling can be incor por at ed" and is assumed to denote "the
a pp r op r ia t e p a t t e r n o f muscular c oor d in at io n stored in a ct ion memory"
( p . 5 52 ). Hence the formula f o r walk amounts simply t o :

(WITH (ATG (ACT))) (x , S, FEET) & CAUSE (S, (on (TRAVEL)) (x, LAND))

To make the d e f i n i t i o n maximally g e n e r a l , a proposal i s added t o the

effect that 'on land' can be removed. The actual ju stificatio n
provided i s t h a t we have no t r o u b l e understanding "walking on a i r " or
walking on wat er (ib id .). Consequently t h e r e i s no s p e c i f i c a t i o n
of direc tio n a l orientation and no i m p l i c a t i o n s as t o destination,
c o n t i n u i t y and r e g u l a r i t y o f motion. The possi bl e r e l a t i o n o f manner
specifying verbs with d irectionality will be discussed in some
de tail, following a b r i e f c o n si d e r at i o n of other c h ara c te ris tic s of
some of these verbs. At this point, suffice it to notice that
d e f i n i t i o n s such as the one j u s t presented, besides being incomplete,
also mix general w it h prototypical information. It is not obvious

- 243 -
why ' f e e t ' , f o r i n st a n c e , should be included and 'on l and' need not;
one can perfectly understand ' wa l k i n g on one's hands' and such a
s i t u a t i o n is perhaps l ess u n l i k e l y than ' wa lk i ng on w a t e r ' .

The in e vita b ility o f c onc ent ra t ing on the t y p i c a l i nstances of the

occurrence of such verbs is also apparent in analyses which are
compl etel y u n r el a t ed t o Prototype theor y and are in essence c o n t r a s t -
based in keeping with the structuralist tradition. Consider, as a
further example, Nida ( 1 9 7 5 :7 3 - 82 ) who s t a r t s wi th the preliminary
r e s t r i c t i o n t h a t hi s a n al y s i s should be a p p l i c a b l e t o the movement o f
persons. Nida i s in search of d i a g n os t i c f e a t u r e s o f the meaning o f
a small set of semantically related verbs such as 'ru n ', 'w alk',
'hop', 's kip ', 'jump', 'dance' and 'craw l'. The resulting
p r e s e n t a t i o n o f t he c on tr a st s e x h i b i t e d by these verbs is i n the form
o f an ext remel y d e t a i l e d m a t r i x i n v ol v i ng the natur e o f the contact
o f limbs w i t h the ground and f e a t u r e s such as number of limbs used,
or der o f c o nt a c t , whether one or another limb i s always in contact or
no limb i s in c ont act a t a l l at ti mes.

The extreme detail of the description is its e lf indicative of the

fact that specific instances r a t h e r than a maximally general picture
of the motions in question are being analysed plus the additional
r equirement t h a t c o n t r a s t s be brought about a t every s i n g l e poi nt and
at the cost of including i nfor ma ti on which is irrelevant for the
understanding and c o r r e c t use o f the items in quest ion. Once again
mixing general with specific i n for ma ti on seems inevitable. Not ic e,
for i ns t an c e, that the specification of 'run' i n vol ve s: '2 limbs
us ed ' , 'a lte rn ativ e or der of contact wi th ground' and 'no limb in
c onta ct at times' which is the only po i nt on the matrix
differentiating 'run' from 'w alk'. The restriction to persons,
however, n e c e s s i t a t e s a s p e c i f i c a t i o n o f '2 l i m b s ' r a t h e r than a more
general one i n v o l v i n g 'any limbs which are normally in c on ta ct wi th
ground'.

Structuralist analyses y i e l d b e t t e r r e s u l t s when t hey compare verbs

which c o n t r a s t more d r a m a t i c a l l y , e.g. 'run' and ' w a l k ' compared to
'ro ll' and 'creep'. The main c ontr ast s in terms of perceptual

- 244 -
properties can be directly borrowed from such analyses for the
description of sim ilar MGMVs and construed as condi ti ons of
application. Examples w i l l be o f f e r e d however to show t h a t not a l l
p r o p e r t i e s have a c o n t r a s t i v e val ue f o r a whole set o f verbs and t h a t
some of the non-contrastive ones are quite relevant for the
understanding o f the items in quest ion.

Change-of-location, c o n t i n u i t y o f the motion and c onta ct wi th ground

are common characteristics of 'trexo' (run), 'p e r p a t a o ' (walk),
'k ila o j' (ro ll) and 'sernome' (creep). The p oi nt s of c o n tr a st
i n vol ve ' conti nuous contact' for the two last verbs rather than
' i n t e r m i t t e n t c o n t a c t ' which c h a r a c t e r i z e s the f i r s t p a i r and 'whole
body/main part o f the body in contact' r a t h e r than 'extremities of
limbs (fee t)'. A pr ope rt y such as 'continuous series of poi nts in
contact w it h ground' juxtaposed to 'any p a r t / p o r t i o n of the body in
contact' can also be evoked in a structu ralist a n al y s i s to
d i f f e r e n t i a t e between ' k i l a o j ' (ro ll) and ' s e r n o m e ' (creep).

Notice, however, t h a t a l o t more i nfor mati on i s necessary which need

not be c o n t r a s t i v e even w i t h i n t h i s small subset. The verb 'k ila o j'

(ro ll) i s s ca rc el y a p p l i c a b l e t o human body motion, as i t r e q ui r es a

round o bj e c t moving and i mpl ie s 'smooth' motion, at least in its
prototypical uses. It is typically used f o r the motion o f a ball,
for i ns tan ce , and in case the o bj e c t is not perfectly round and
smooth ( e . g . a barrel r a t h e r than a b a l l ) the verb also invokes the
p i c t u r e o f a r e c l i n i n g sur face or obvious e xt e rna l i n s t i g a t i o n o f the
motion. A further p r op e r t y , which characterizes this motion and
under which t he c o n d i ti o n of 'smoothness' can be subsumed, is
're g u larity ', which seems relevant to a number of verbs of this
domain. Points o f c o n t r a s t between 'k i l a o j ' (ro ll) and ' k a t r a k i l a o '
(roll down) include ' r e g u l a r i t y ' and 'smoothness' which are absent in
t he l a t t e r verb. Besides these p r o p e r t i e s , 'katrakilao' (roll down)
also i nvolves 'downward d i r e c t i o n ' , a p p l i c a b i l i t y to human body, and
typically 'loss of equilibrium ' and ' g r a v i t y ' as the cause o f motion.
'Continuous c ont act w it h ground' is, as al re ady mentioned, a shared
pr ope rty of these two verbs w it h ' sernome' (creep). 'Body
involvement' as a sol e means o f c h a n g e - o f - l o c a t i o n can be considered
as a point of contrast between 'sernome' and t he aforementioned

- 245 -
verbs. Other properties, present at least in its prototypical
i ns tance s, such as 're lativ e ly slow motion' and CL which is not
ty pical for humans, need not , however, be mentioned in connection
with every o t he r verb of the set. F ur the r properties w ill be
discussed in connection w it h the test results in the following
c hapt er .

In Chapter 2 on the S-P-E d i s t i n c t i o n , a basic d i f f e r e n c e was noticed

between the verbs j u s t discussed and ' p i b a o ' (j ump). In simple terms
this can be s ta te d as a c o nd i ti o n t h a t ' p i b a o ' i nvolves 'momentary'
loss of contact w i t h ground. A f u r t h e r p oi n t of contrast could be
that 'pibao' (jump) i s l i n k e d wi th 'upward' direction, a pr ope rty i t
shares wi t h 'x o r o p i b a o ' (hop). The motion described by 'pibao' can
be a 'punctual occurrence' or an a c t i v i t y ( s e r i e s o f repeated jumps),
w h i le 'xoropibao' (hop) is n e c e s s a r i l y c h a r a c t e r i z e d by ' c o n t i n u i t y ' .
An important prototypical c o ndi ti on for the application of both
'pibao' (jump) and 'xoropibao' (hop) i s t h a t they are not normally
used f o r covering a d i st an ce in order to reach some d e s t i n a t i o n on a
non-vertical axis. Thi s c o nd i t i on seems t o be a more prominent poi nt
o f c o n t r a s t wi th the ot he r verbs alr eady discussed, but i s completely
absent in the literature as far as I know. I n c l ud i n g 'order of
contact' in the specification of these verbs in or der to c on tr a st
them (as Nida 1975 does) would be mi sl eading: no p a r t i c u l a r order of
c onta ct seems t o be c h a r a c t e r i s t i c even o f the p r o t o t y p i c a l uses o f
'xoropibao' (hop) . The boundaries between 'pibao' and 'xoropibao'

are e v i d e n t l y f u z z y . Concentrati ng on t h e i r r e s p e c t i v e p r o t o t y p i c a l
c ondi ti ons of application can bri ng out the essential poi nts of
contrast which seem to be related to 'seri ousness of purpose',
' s i n g l e vs repeated motion' and ' r e g u l a r i t y o f repeated jumps/hops'.
It seems plausible to assume t h a t prototypical conditions for the
application of 'pibao' (jump) i nv ol ve moving over an obs ta cl e in
order to continue on one's course, t r y i n g t o reach something high, or
physical e xe r ci se (related to regular and repeated such moti ons).
Compared to the prototypical instances of 'x o r o p i b a o ' (hop) which
i n vo lv e 'motion f o r recreational purposes', those o f 'pibao' (jump)
are 'marked' for 'seri ousness of purpose'. The p o i nt s of overl ap
(i.e the 'focus' of fuzziness) become evident; when continuity/

- 246 -
r e p e t i t i o n and r e g u l a r i t y are present but (seriousness o f ) purpose i s
in doubt, choice between the two verbs in question w i l l be random.

Another 'manner' s p e c i f y i n g verb which can be only m a l t r e a t e d w i t h i n

the framework of c o nt ra st - ba se d theories is 'xorevo' (dance).
Regularity of motion, nature of contact wi th ground, poi nts of
contact e t c . seem compl etel y i r r e l e v a n t in t h i s case. Nida ( 19 75:75)
provides a value f o r a l l the dimensions r e l e v a n t f o r 'walk' or ' r u n '
in analysing 'dance', w h i le M iller and Johnson-Laird (1976:551)
consider it an i nstance of a 'specialized mode' of "travelling by
f o o t on l and", in the same subset wi th 'hop' and ' s k i p ' . It is hard
to see the relevance o f such c onsi de ra ti ons f o r the average speaker' s
competence with re spe ct to this ext remel y idiosyncratic verb.
Although i t i mpl ie s 'continuous m o ti o n ' , it is not used f o r covering
di st an ce or reaching a destination but is clearly marked for
'purpose' (loosely specified here as 're c re a tio n a l'). Stating th a t
the motion i nvolved is 'rhythmic' does not amount to much e i t h e r ;
typical (and o f t e n m a r gi n al ) i nstances o f i t s a p p l i c a t i o n in terms o f
perceptual properties can be only described in connection with
s p ec ific cultures, a s i t u a t i o n not l i k e l y t o a r i s e i n any o t he r kind
o f movement discussed so f a r . R e s t r i c t i o n to humans i s also p a r t o f
its specification; the dancing o f t r a i n e d bears and c o u r ts h i p dances
of b i rds can be s a f e l y exclu ded even from general (as opposed to
prototypical) d e fin itio n s , as extensions o f meaning.

Two more 'bodily movement' verbs w ill be discussed very briefly,

which necessitate a specification of 'medium' and are mentioned
earlier in th is s ec t ion in that connection, namely 'k o lib a o ' (swim)
and 'p e t a o j ' (fly ). Not ice t h a t a specification o f limbs used, for
i ns t an ce , is completely irrelevant for either of them. Nida
(19 75 :7 9) i ncludes 'forelimbs used as a means of propulsion' for
'fly ', f o r the sole purpose ( i t seems t o me) o f c o n t r a s t i n g i t wi th
ot he r b o di l y movement verbs at every s i n g l e p o i n t . E v i d e n t l y 'wings'
and not ' f o r e l i m b s ' ( i n g e ne r a l ) needs to go i n t o its specification.
N ot ic e , a l so , t h a t 'kolibao' (swim) i s not mainl y used f o r reaching a
d e s t i n a t i o n and t h a t i t s d i r e c t i o n is t y p i c a l l y h o r i z o n t a l e s p e c i a l l y
i f appl ie d t o humans.

- 247 -
General definitions and contr ast - ba se d matr ices cannot r eveal these
properties, which seem, however, central t o the understanding o f the
verbs in quest ion. In f a c t , the r e l a t i o n s between 'd ire c tio n a lity '
and 'manner' are r a r e l y (if at a ll) taken into account, except f o r
v er y obvious cases like that of 'clim b' or 'jump'. It seems,
however, that different 'manner' specifying motion verbs exhibit
different degrees of c o m p a t i b i l i t y wi th vertical and h o r i z o n t a l
direction.

It seems p l a u s i b l e to assume t h a t 'k o l i b a o ' (swim) a p p l ie d to human

motion i s the most c h a r a c t e r i s t i c case o f h o r i z o n t a l d i r e c t i o n since
t he motion i t descri bes i s t y p i c a l l y executed on the surface o f the
sea, i.e. on n e a r l y zero g r a d i e n t . Swimming towards the bottom of
t he sea or from the bottom to the surface i s l ess t y p i c a l and in most
such cases verbs marked f o r v e r t i c a l i t y w ill be used i n st e a d , i.e.
' k a t a b i o m e ' (dive) and ' a n a b i o m e ' (emerge ( out o f w a t e r ) ) . Motion in
a ll different directions within wat er as in scuba diving, for
i n s t an c e, arguabl y i nvolves marginal i nstances of t he verb's
application.

Motion in the air, i.e. in (three di mensi onal ) space expressed by

'p e t a o j' (fly ) is c h a r a c t e r i s t i c a l l y i nde te rmi na te as to d i r e c t i o n a l
orientation. In practical terms t h i s means t h a t only completely
vertical direction may be s p e c i f i e d wi th an expression marked for
vertic a lity, i.e. 'kano vutja' (dive). In terms o f prototypical
c on di ti o ns o f a p p l i c a t i o n 'petaoj' can be sai d to be compatible w i th
a variety of d iffe r e n t directions, i.e. not s p e c i f i e d f o r 'ty p ic ally
horizontal d i r e c t i o n ' unlike ' k o l i b a o ' (swim).

Compared t o these two verbs, ' p e r p a t a o ' (walk) i s an in between case,

i.e. much l ess flexible than 'petao j' (fly ) but more so than
'kolibao' (swim). If walking is excecuted on q u i t e a precipitous
grade, a verb marked f o r v e rtic a lity is l i k e l y t o be used i n st e a d ,
i.e. 'a nifo rizo ' (go u p hill). Compared to 'p e r p a t a o ' (walk),
' sernome' (cr eep, c r aw l ) is more r e s t r i c t e d to horizontal direction
for obvious p h y s i o l o g i c a l reasons, i . e . owing to the physical
ab ilitie s of the human body. A very p r e c i p i t o u s grade w ill

- 248 -
necessitate use of 'ska rfa lo n o ' (cl imb) i n st ea d. More accurately
put, i t w ill push in the d i r e c t i o n o f ' s k a r f a l o n o ' .

In a language such as MG which does not have separate lexemes for

'creep' and ' c r a w l ' , the boundaries between 'sernome' (c re ep, cr awl )
and ' s k a r f a l o n o ' ( c l i mb ) are expectedl y fuzzy and choice between the
two verbs is bound t o depend on how slopy the surface i s , in marginal
cases of the use of either of them. The situation is clearly
d i f f e r e n t in the case o f moving o bj ect s with physical a b i l i t i e s other
than those of humans, e.g. snakes, where 'sernome' ( cr eep) w ill
p r e d i c t a b l y cover a wi der area than ' s k a r f a l o n o ' (climb). The l a t t e r
verb w i l l be only used in such cases, i f the d i r e c t i o n is completely
vertical (e.g. a snake going up a t r e e ) .

Gradation is also apparent in the case of verbs prototypically

understood as i n v o l v i n g 'downward' d i r e c t i o n . In a maximally general
definition, 'ylistra o ' (slip, slide) can be only u n s p ec if ie d as to
v ertic a lity, but in its prototypical uses, it certainly r e qu i r e s a
c o n di ti o n of 'downward d i r e c t i o n ' . Compared to 'y7i s t r a o j ' (slip,
slide), 'kila o j' (ro ll) seems to be less marked for 'downward
direction'. In terms o f p r o t o t y p i c a l i t y c o n d i ti o n s , t h i s means t h a t
'ylistrao ' is a central case o f ' p r o t o t y p i c a l l y downward d i r e c t i o n ' ,
w h i le 'kilao' is l ess so. Clearly, if the surface is completely
vertical, n e i t h e r verb w i l l be a p p l i c a b l e any more. The choice w i l l
be r e s t r i c t e d to 'pefto' (fa ll), 'katrakilao' (roll down) or some
ot he r 'downward' motion verb implying e i t h e r 'inte rm itte n t c ont a ct '
wi th surface or ' l o s s of c o n t a c t ' wi th supporting s u rf ace . Thus, if
one imagines a line l e adi ng from a completely horizontal to a
competely v e r t i c a l direction, poi nts along t h i s l i n e can be occupied
by vari ous 'manner' s p e c i f y i n g verbs in the f o l l o w i n g sequence:

'kolibao' ' sernome' 'perpatao' 'petaoj' 'kilao' 'y lis tra o '
HORIZONTAL VERTICAL

A possi bl e schematic representation of the relations of

'd irectio n ality' and 'manner of motion' exhibited in some verbs
i n v o l v i n g 'human body c onta ct w it h ground' i s the f o l l o w i n g :

- 249 -
 HORIZONTAL

/ Oo
/ ^ <XJ
/ —

^ f5 <+-
£ £ O-

VERTICAL

Clearly, t he above b r i e f p re se n t a t i o n only touches on the problem o f

combinability between horizontal-vertical direction and 'manner' of
motion. It is hoped, however, t h a t such an approach can be used and
expanded so as to cover most 'manner' s p e c i f y i n g MVs as wel l as most
' g e n e r a l ' motion ones.

4.3 Hi e ra rc h y o f p r o p e r t i e s : taxonomies and paradigms

It can be easily shown that almost any h i er a r c h y which could be

proposed as hol di ng between 'maj or properties' would be a r b i t r a r y .
I t has also been argued here t h a t c a t e g o r i z a t i o n o f the same m a t e r i a l
depends h e a v i l y on the angle from which the a n al y st wishes t o look a t
it and the s o r t o f f e a t u r e s s/he decides to use. Thi s also i mpl ies
that hierarchical structuring between 'minor properties' is also
a rbritrary, at least to some extent. The e xt e n t to which these
claims are true can be discussed at th is point, since most o f the
relevant properties have been presented, analysed and e x e m p l i f i e d .
The purpose o f t h i s di scussion is t wo fo l d. It should f i r s t show t h a t

- 250 -
in s t r u c t u r a l i s t terms the f i e l d under i n v e s t i g a t i o n i s probably what
Lounsbury ( 1 9 6 4 : 1 08 6- 7 ) would consider 'neither a genuine taxonomy
nor a genuine paradigm' but closer t o the l a t t e r rather than the
former, unless no distinction is drawn between ' dimensions' and
'features'. It should secondly l ead to the conclusion that
d i s cov er ing the 're la tiv e salience' of properties is a more
wor thwhi le t a s k than di s co v er in g t h e i r r e l a t i v e h i e r a r c h y , since the
former but not the l a t t e r can come out of s p e c i f i c tests e licitin g
i n fo rm at io n from native speakers and giving, therefore, a less
arbitrary picture. A related i ssue, which is also more i n t e r e s t i n g
than hierarchy of properties, is the 'non-arbitrariness' of their
combinations, which w i l l be taken up in the f o l l o w i n g s e c t i o n .

In 4.1 it is argued t h a t 'major' properties are o f t e n linked with

d i f f e r e n t p r e d i ca t i o n s and d i f f e r e n t forms o f the same verb, so t h a t
both ' s t a t i v e ' and ' n o n - s t a t i v e ' , f o r i n st an ce , may be a p p l i c a b l e to
' a n e v e n o ' (ascend). This w i l l depend, among ot he r t h i n g s , on whether
it appears in the Perf.b' form as '/me anevasmenos' , or in the
Pres.If, form as 'aneveno'. Even i f we ignore such problems (along
wi th the phenomenon o f g r a d a t i o n ) and account only f o r a p a r t i c u l a r
'princip al' form o f a ver b, i.e. the P r e s . I f . one, a s t r i c t h i er ar chy
seems impossi bl e, especially i f we also do away wi th the major-minor
features distinction adopted in the present study. Not ic e a number
o f p o s s i b i l i t i e s f o r such f e a t u r e s in the area under i n v e s t i g a t i o n .

Anderson (1971) envisages a p a t t e r n o f the form:

STATIVE NON-STATIVE

NON-AGENTIVE AGENTIVE

CAUSATIVE NON-CAUSATIVE

- 251 -
I nstead o f the above p l a u s i b l e schema, one could also have:

POSITIONAL (vs NON POSITIONAL: e . g . be vs t h i n k / h a v e )

DYNAMIC (Move)

CAUSATIVE (Cause) NON-CAUSATIVE (Non-cause)

CAUSE CAUSE TRANSITIVE INTRANSITIVE

CHANGE- CHANGE-
OF-LOCATION OF-POSITION

A lternatively, CAUSATIVE could branch i n to AGENTIVE and NON-AGENTIVE

and i nstead o f CAUSE CHANGE-OF-POSITION above, one could have NON­
CAUSE CHANGE-OF-LOCATION (sticking to genuine structuralist
contrasts). A pos si bl e branching o f t h i s new category could be:

NON-CAUSE
CHANGE-OF-LOCATION

CAUSE NOT CAUSE CAUSE

CHANGE-OF-POSTURE CHANGE-OF-POSTURE CHANGE-OF-ORIENTATION

The possibilities are not exhausted. As a f i n a l example consider

a lso:

- 252 -
 TRANSITIVE

MOTIONAL NON-MOTIONAL

[-CAUSATIVE CAUSATIVE

CHANGE OF CHANGE OF
ORIENTATION POSITION CHANGE OF
(ROTATION) LOCATION
general manner regular/
repeated

general manner impetus random walk

INTRANSITIVE

MOTIONAL NON-MOTIONAL

CHANGE OF CHANGE OF CHANGE OF POSITIONAL NON-POSITIONAL

LOCATION POSITION ORIENTATION

It is fa irly un c on tr o ve rs i al that at least some e n t a il m e nt

relationships can be de te ct ed as existing between the properties
under consideration. Both CL and CP, for i n st a n c e , do imply
'm o tio n al)' and this order seems irreversible. Th e re fo r e some
taxonomic s t r u c t u r i n g a t i n t e r m e d i a t e l e v e l s can be envisaged.

In 4 .1 i t i s also argued t h a t ' g e n e r a l ' means simply 'not specifying

manner' and does not suggest a hierarchical connection between the
two, so such a p o s s i b i l i t y is r ul ed out. This involves, however, a
broad understanding of 'manner' as presented in 4.2.4. Different
understandings would obvi ousl y lead to different classifications.
N ot i c e , further, that in 4.2.1 'g eneral' CL is linked to
'd ire c tio n a lity '. Where directional features stand in connection

- 253 -
w ith 'g eneral' and 'manner' specifications is, however, an open
quest ion. It i s a c t u a l l y suggested in 4 . 2 . 4 t h a t ' d i r e c t i o n a l i t y ' is
also linked with 'ma nne r' . Whether this is accepted or n o t, the
re latio n /relative hierarchical order o f CL, CP and 'd irectio n ality'
allows for a number of possibilities not discussed y e t . Although
only CL i s usually analysed as i n v o l v i n g d i r e c t i o n a l properties, it
can be shown t h a t CP can also subsume such components. Consider,
firs t, some r e l a t i o n s between CL and ' d i r e c t i o n a l i t y ' which have been
a lr e ad y hi nted a t in previous s ec t i o ns :

CHANGE OF LOCATION

DIRECTIONAL < -------------------------------------> DIRECTIONAL

vertical (horizontal) i n d et e r mi n a t e \ \
o r i g i n and path dependent
destin ation

As a lr e a d y poi nted out in 4 . 2 . 2 , 'accompaniment' could appear under

'dependent' motion, or be understood i n Gruber's (1965, 1976) sense,
in which case i t is a p r o pe r ty o f much hi gher l e v e l i n v o l v i n g goa l,
d irectionality, etc. 'D irectionality' can also appear under CP.
Verbs such as '/csap/ono^' (lie down), 'skivo' (bend), 'sorjazome'
(collapse), i n v ol v e 'downward direction' in every possi bl e use.
Besides, 'kaQome^' (sit down) is prototypically understood as
i n v o l v i n g 'downward d i r e c t i o n ' , i.e. s i t t i n g from a standi ng p os i t i o n
and l ess t y p i c a l l y as i n v o l v i n g 'upward d i r e c t i o n ' , i.e. sittin g up
from a l y i n g p o s i t i o n . A d i f f e r e n t verb ' a n a k a Qo me ' ( s i t up) can be
used in the l a t t e r case, which obvi ousl y i nvolves 'upward d i r e c t i o n ' ,
as does also 'a n ap /b a o' (jump up). All these verbs are c l a s s i f i e d
under CP in L i s t V I I and i n p a r t i c u l a r , some o f them are under 'CP
resulting in d i f f e r e n t p o s tu r e' w h i le others appear under 'CP non­
resulting in d i f f e r e n t posture'. T he ref or e 'd irectional' could be
above CL and CP.

- 254 -
In the previous sec ti on it is argued that 'change of orientation'
involving ' r o t a t i o n ' could be e i t h e r on the same l e v e l wi th CL and CP
or under both as i t involves verbs un s pe ci fi e d f o r CL or CP. It is
also suggested t h a t ' i n s t r u m e n t a l i t y ' may be subsumed under 'manner'.
Especially if no distinction is drawn between ' dimensions' and
'fea tu res', a number o f d i f f e r e n t possibilities exist, with respect
to 'instrum entality' and 'medium'. They can be s ch e ma t i ca l l y
r epresented as f o l l o w s :

MED INSTRUMENT

LAND WATER AIR BODY CONVEYANCE GRAVITY

INSTRUMENT

BODY CONVEYANCE LAND WATER AIR

LIMBS WHOLE BODY

FEET FEET & HANDS

These schemata are a b br ev i a t i o n s o f 'compl et e' versi ons in which most

f e a t u r e s under ' i n s t r u m e n t ' would appear under each separ ate 'medium'
specification and v i c e versa. What i s more i m por ta nt , 'instrument'
links up w it h 'cause', so t h a t a confusion between hi ghe r and lower
level p r o p e r t i e s i s again possi bl e ( i g no ri n g the p r i n c i p l e s suggested
here as relevant for the d istinction). Never the le ss , taxonomic
relations can be again de t ec t ed , wi th the proviso t h a t 'dimensions'
and ' f e a t u r e s ' are not kept separ ate. Not ic e, f o r i n st a n c e , t h a t the
hiera rchical order o f whatever appears under BODY in the above schema
cannot be r ever sed. N o t i c e , f u r t h e r , t h a t f e a t u r e s i n v o l v i n g 'number
of limbs used' or different types of 'contact w it h ground', e.g.
'continuous', 'i n t e r m i t t e n t ', entail obviously ' c o n t a c t w i t h ground',
but 'loss of contact' would then have to appear on a hi gher level
than, say, 'conti nuous c o n t a c t ' and i n p a r t i c u l a r on the same l e v e l
w i t h ' c o n t a c t w i t h ground' .

- 255 -
P r o p er t i e s such as 'distance cove ri ng' are also necessarily lower
than CL, and '+ /- fast motion' is lower than 'manner' in any
understanding of the latter. In 4.1.3 a number o f properties are
mentioned which are relevant to specific verb taxonomies, such as
those proposed for hyponyms of 'fev\o' (leave), e.g. 'negative
connotation', ' going away from a group o f p e o p l e ' , 'changing country
of residence', etc. There can be l i t t l e doubt t h a t these p r o p e r t i e s
are hierarchically lower than CL, 'distance covering', and
'd ire c tio n a lity ', for i n s t an c e. It is not therefore true that
features cannot be used to support the attested hierarchical
structure between hy pony mi ca ll y/ ta xonomi ca ll y related items (Rhodes
1983). N i d a ' s taxonomy (discussed in 4 . 1 ) i s not convi nci ng because
it i nvolves terms which are probably on the same h i e r a r c h i c a l level.
If 'proper' taxonomies are e st a b l i s h e d (and i t is suggested in 4 . 1 . 3
that they can be onl y established at a fa irly low level in this
field ), f e a t u r e s d i s t i n g u i s h i n g lower from hi gher l e v e l t a x a ( i n the
same taxonomy) do support the hierarchical structure, almost by
d efinition. The problem seems t o be whether t he a t t e s t e d d i f f e r e n c e s
between a basic l e v e l category and i t s hyponyms can be construed in
the form of 'tra d itio n a l' features, i.e. undescriptive (if not
primitive) one/two-word terms or not. It is suggested here that
except f o r few cases the answer i s ne ga t iv e.

It is further suggested that the properties in question have a

specific hierarchical s ta t us only w i t h i n specific taxonomies. This
means t h a t there is no a p r i o r i reason why 'negative c onnot at ion'
should be lower than ' t r a n s i t i v i t y ' or ' n o n - c a u s a t i v e ' , f o r i ns tance,
except for t he major-minor properties distinction proposed here.
Wi thi n taxonomy 5 ( L i s t V I ) , ' n e g a t i v e connota ti on' i s used f o r verbs
such as 'ksekubizome' , 's t r i v o ' and the like , all equivalent to
'clear o f f ' . In t h a t p a r t i c u l a r taxonomy, it can be said t o be even
l ower l e v e l than some p r op e rt y : ' o t h e r people i n v o l v e d ' , a lso present
in ot he r hyponyms o f ' f e v y o ' ( l e a v e ) such as ' a p o x o r o ' (withdraw) and
'aposirome' (retire). Wi thi n taxonomy 11 ( L i s t V I ) , i t could be used
to characterize verbs such as 'katrakilao' and 'kutruvalao' (roll
down) and i t is f a i r l y c le a r th a t, in t h a t case, i t has nothing to do
w ith ' o t h e r people involved'. Such consi de ra ti ons do not, however,
bear on how s a l i e n t the p r o pe r ty of 'negative c o n no ta t io n' may be

- 256 -
psychologically. In the next c hapt er di scussing test results, its
relative salience can be p a r t l y assessed in connection w it h other
properties, traditionally considered relevant for MVs such as
'd irectio n ality' and ' c o n t a c t w i t h ground'. For the moment, a l l that
can be said is that its exact hierarchical position cannot be
independentl y determined.

In traditional structuralist terms it must be noti ce d that

paradi gmati c structuring is also exhibited in the field under
investigation. Paradigms are e xe m p l i f i e d by the p a i r s :

'beno' (go i n ) - 'vyeno' (go o u t ) ,

'vazo' (put i n ) - 'vyazo' ( t ak e o u t ) ,

' i s a y o ' ( im p o r t ) - ' e k s a y o ' (export)

'iserxome' (enter) - 'ekserxome' (exit)

where only IN-OUT are c o n tr a st e d . S i m i l a r par adi gmati c r e l a t i o n s are

e x h i b i t e d in the f o l l o w i n g p a i r s :

'aneveno' (go up) - ' k a t e v e n o ' (go down)

'anevazo' ( t ak e up) - ' k a t e v a z o ' ( t ak e down)
' a n i f o r i z o ' (go uphill) - 'katiforizo' (go d o w n h i l l )
' ip s o n o ' (raise) - 'xamilono' (l owe r)
'anerxome' (ascend) - 'katerxome' (descend)

D e i c t i c p a i r s e x h i b i t also paradi gmati c r e l a t i o n s :

'piyenoj' (go) - 'erxome' (come)

/p iy e n o 2 / (take to) - 'fer/?o' (bring)

It seems, however, t h a t 'fevyo' (leave) can be a lso c ont ra st ed to

' e r x o m e ' (come) and ' s t e l n o ' (send) to ' f e r n o ' (bring), although more
than a ' s i n g l e f e a t u r e c o n t r a s t ' i s i n v o lv e d.

- 257 -
The psychological validity of some o f these p a tt e r n s and p r o p e r t i e s
w ill be discussed in the f o l l o w i n g chapt er.

4.4 Motion verbs and the n o n - a r b i t r a r i n e s s of c at e gor i es

Mervis and Rosch ( 1 9 8 1 : 9 1 - 2 ) present evidence to show t h a t a t t r i b u t e s

(p roperties) are not combined arb itrarily to form items.7 if the
opposite were true a ll combinations of attribute values would be
e q u a l l y l i k e l y to occur. Mervis and Rosch's ( i b i d . ) illu stratio n of
this issue i nv ol ve s the properties normally used in classifying
animals, i.e. "coat" (fur, feathers), "oral opening" (mouth, beak)
and "primary mode o f locomotion" (flying, on f o o t ) . If animals were
created according to a model contending that the division of re al
world o bj e ct s into c at e g o r i e s is originally arbitrary, one would
expect e i g h t d i f f e r e n t t ypes/ combi nat ions, e . g . :

animals w it h f u r and mouths moving p r i m a r i l y on f o o t

animals w i t h f u r and mouths moving p r i m a r i l y by f l y i n g
animals w it h f u r and beaks moving p r i m a r i l y on f o o t
etc.

It is fa irly evident that the perceived world of ob j e c t s is not

actually structured in this manner. Only two of t he eight
t h e o r e t i c a l l y p o s si b l e combinations o f a t t r i b u t e values "comprise the
great majority of existent species in the world that are possi bl e
based on t h i s total set " (ib id .), i.e. mammals (fur, mouth, feet),
and bi rds (feathers, beak, flying). It would, in fact, be more
accurate to say that c r ea tur es with f e at h e r s and mouths moving
primarily on foot are not likely (if at a ll) to occur and that
c r ea tur es w it h f e a t h e r s and beaks moving p r i m a r i l y on f o o t , although
they do occur ( e . g . chicken), are q u i t e expectedl y judged by subjects
as marginal i nstances o f the relevant super or di nate c at egor y, i.e.
'b ird '.

In the area o f MVs t h e 'non-arbitrariness' o f c a t eg or i e s i s e q ua l l y

(if not more) obvious, although i t is perhaps o f a d i f f e r e n t na tur e.
Properties such as ' in te n tio n a lity', ' a n i ma c y ' , 'presence of

- 258 -
supporting s u r f a c e ' , ' l i q u i d moving o b j e c t ' , etc. w i l l combine or not
in an even less arbitrary and qualitatively different way from
attributes observed in d i s c r e t e o b j e c t s . Combinations in t h i s field
depend, to a large extent, on the laws of physi cs, besides the
na t ur a l characteristics of moving objects, e.g. the na tur al
potentials of human beings as moving organisms. The issue is both
broad and complex, so that tackling it in any detail would lead
beyond the scope o f the present investigation. It seems, however,
worthwhi le p o i n t i n g out a number o f s e l f - e x p l a n a t o r y cases in support
of Rosch's principle from an area of the vocabulary which is
compl etel y different from the ones Rosch and her col leagues have
d e a l t w i t h so f a r . The p r i n c i p l e seems to me t o gi ve a new prospect
to lexical semantic analysis, which has so far handled relations
between c o - e x i s t i n g p r o p e r t i e s in the form o f e n t a i l m e n t s .

Consider as an example 'i mpe tus ' as a c h a r a c t e r i s t i c o f a number o f

causati ves o f motion such as ' pet ac> 2 ' ( t h r o w ) , 'eksfenbonizo' (hurl).
Presence o f ' i mpe tus ' can be said to imply 'absence o f accompaniment'
or 'loss of contact'. This relation can be handled through
e n t a il m e n t or semantic lexical redundancy rules (although this
practice is not a p p li ed in the t reatments o f MVs t h a t I know o f ) .
The combination of 'i mpe tus ' with 'loss of contact' may in other
words be considered a n e c e s s a r y / i n e v i t a b l e one. The same appl ie s to
the r e l a t i o n of ' i mp e tu s ' w it h some pr ope rt y we could c a l l 'speed'/
'energy'. Its combination w i t h 'travel through a i r ' is, however, a
l ess necessary one, since obj ect s can be 'h urled' while in wat er ,
from the sur face of wat er into it, etc. Wi thi n the framework o f
pr otot ype t heor y 'travel through a i r ' should f e a t u r e as p a r t o f the
c ombi nat or ial possibilities of 'impetus'. Wi thi n the framework o f
necessary and s u f f i c i e n t c ondi ti ons approaches 'travel through a i r '
appears u s u a l l y as a separate p r o pe r ty , unr el at ed to ' i m p e t u s ' .

The p ro pe r ty of ' c on ti n uo us ' motion is clearly combinable wi th

'r e g u la r/re p e a te d ' or the notion o f 'random wal k' as e xe m p l i f i ed in
' p e r i p i anjerne' (wander) and the l i k e .

- 259 -
'Liquid'/'m ass' as type of moving o bj e ct co-occurs normall y wi th
'downward' rather than 'upward' direction, unless 'i mpe tus ' is
invol ve d. The l a t t e r p o s s i b i l i t y does e x i s t but i s somehow marked/
ma rg ina l. The p oi n t here is t h a t verbs r e s t r i c t e d to d e sc r ib i ng the
motion o f liquids onl y, e.g. 'stazo' (drip), 'x i n o m e ' (be spilled)
are more likely than not to also i nv ol ve 'downward' motion. An
apparent exception t o t h i s i s provided by the verb 'atiavlizo' (gush
out) which, however, may be also thought to involve 'im petu s'/
'e n er gy ' and can t h e r e f o r e h a r d l y be c a l l e d an e xc ept ion. This verb
is an example o f the co-occurrence of 'upward' and ' out war d' which
although f a r from necessary, it i s c e r t a i n l y much commoner than t h a t
o f 'out ward' and 'downward', e s p e c i a l l y in the case o f ' w a t e r ' as the
medium o f motion ( f o r obvious re as ons ). Characteristic examples of
these combinations are 'anafiiome't which involves ' o u t o f water and
upward' and ' k a t a b i o m e ' which involves ' i n t o water and downward'. It
is also i n t e r e s t i n g t o n o t i c e t h a t although t h er e e x i s t verbs which
combine 'up' and ' o u t o f ' , e.g . 'ksepetayome' (jump out of, appear
suddenly out of), the combination of ' up' and 'in to' does not
m a t e r i a l i z e in any MGMV.

Quite evidently motion into wat er w ill also be 'downward', a

combination e x e m p l i f i e d i n verbs such as ' v i Q i z o m e ' / ' v u l j a z o ' (sink),
'k a t a S i o m e ' / ' v u t a o ' (dive), and the l i k e . A generalization (if not
an e xp l a n a t i o n ) of these simple observations is the obvious
combination of gravity as the cause of motion with 'downward
d irectio n'. I f the moving o b j e c t l acks a s el f- movi ng (he nc ef or th SM)
mechanism, its motion is more likely than not to have 'downward'
rather than 'upward' direction, unless there is an e x t e r n a l cause
i nvolved such as 'impetus', 'accompaniment', etc. In short, '-SM'
combined w it h absence of an obvious e xt er na l cause is naturally
combinable w i t h 'downward' d i r e c t i o n .

The n a tu r al a b ilitie s of the human body being fa irly lim ited to

specific kinds of change of location, a number of theoretically
possi bl e combinations are ruled out . Three examples w ill be
provi ded. Not ice that a specification of 'fe e t' as 'instrument'
(means o f pr op ul s io n ) combines wi th ' intentionality' and 'ground as
supporting surface'. Absence o f 'in te n tio n a lity' does not combine

- 250 -
ea s ily with 'upward d i r e c t i o n ' , but is fa irly common in connection
w i t h 'downward d i r e c t i o n ' , e . g . 'y l i s t r a o ' (slip ), 'k a tra k ila o ' (roll
down), 'p e fto ' (fall (down)), 'g rem izom e' (collapse). 'Dependent
motion' is linked with ' in te n tio n a lity' as t y p i c a l o f animate moving
objects, e s p e c i a l l y in the case of motion 'after' a person, e.g.
' ka ta d io k o ' ( c ha se ) , ' k i n i y a o ' (hunt , run a f t e r ) , ' a k o lu Q o ' (follow).
In g e n e r a l , 'manner' o f motion i s l i n k e d to ' a g e n t i v i t y ' . Not ice the
case of 'p e r p a t a o ' f r e x o ' , ' s k a rfa lo n o \ 'p/Sao', and all their
hyponyms, where the specific 'manner' of motion i nvolved is a
concomitant f a c t o r o f ' a g e n t i v i t y ' .

A specification of ' f e e t and arms' as ' i n s t r u m e n t ' is rarely (if at

all) combinable wi th compl etel y 'downward' direction. In secti on
4.2.4 i t i s argued t h a t the g r a d i e n t o f the i n c l i n e o f the surface on
which the motion is executed lim its the number o f possi bl e
combinations w it h d i f f e r e n t types of human body motion. Some o f the
observati ons made t h e r e can be r e - i n t e r p r e t e d in terms o f the 'non­
arbitrariness of categories'. At this stage, where none of these
observati ons are e la b o r a t e d on, choice between different
i n t e r p r e t a t i o n s or r e p r e s e n t a t i o n s would be premature.

- 261 -
 Notes on Chapter 4

1. If one compares different forms of the same ver b, e.g. 'ime

anevasmeni ' (I am ascended/mounted on) and ' a n e v e n o ' (I am going

up/ ascend), one could say t h a t the former i s ' l e s s m o t i o n a l ' than
the latter (as aready argued in Chapter 2). Participles like
' i me a n e v a s m e n o s - i ' r e c e i v e the most s t a t i v e in terp retatio n that
can be pos si b ly a t t r i b u t e d t o a MV. Besides, it i s po s si b le to
consi der t h a t CL verbs such as ' p i y e n o ' (go) are 'more m ot iona l '
than CP ones involving 'p a rtia l motion' such as 's a le v o ' (move
s lig h tly, s tir).

2. In the f i e l d under i n v e s t i g a t i o n 'csusativity' may be thought of

as a 'c rite ria !' ( r a t h e r than a graded) property; it is mainly
understood here as equivalent to 'motion due to ext er na l
factors'. In Chapter 3 i t is, however, shown t h a t t he issue is
in general (i.e . i r r e s p e c t i v e o f the specific semantic f i e l d ) a
controversial one. It is arguable t h a t 'h it', for i ns t an ce , is
probably 'less causative' than 'k ill', since in the former case
it i s not obvious what the change o f c o nd i ti o n o f the o b j e c t i s .
Wi thi n t he f i e l d o f MVs i t i s shown also t h a t t h e r e i s a t l e a s t a
d i f f e r e n c e in degree o f causation between ' d i r e c t ' and ' i n d i r e c t '
causati ves and t h a t g r a da t i o n may be de tect ed both w i t h i n the
m a ni p ul at iv e and the non-mani pul at ive kind of c au sa ti on. In
s h or t , t h e r e i s good reason to b e l i e v e t h a t 'causativity' its elf
should be regarded as graded.

3. I am not saying t h a t Nida i s suggesting any such a b s u r d i t y . I am

simply t r y i n g to show what the r e s u l t s would be if we were to
push t hi ngs t o t h e i r l o g i c a l conclusions.

4. The l e x i c o n o f MG bears witness to a ' d i g l o s s i c ' sociolinguistic

past (see Ferguson 1959, Kazazis 1982). Present day Greek as
used by most speakers i nvolves a l a r g e number o f 'katha re vous a'
elements which cannot be ignored. Kazazis (1 98 2: 1 11 ) observes

- 262 -
 t h a t many forms which were o r i g i n a l l y 'k a tha re v ous a' are " pa r t o f
the v e r n a c u l a r o f some, and sometimes most, speakers o f Greek",
and t h a t the combinations o f o r i g i n a l l y ' katharevousa' elements
present in the speech o f educated Greeks d i f f e r from one i d i o l e c t
to another . What constitutes standard 'd im o tiki', in ot he r
words the standard language spoken in l a r g e c i t i e s ( or 'Koine Nea
E llin ik i') is s t i l l a ma tt er of great debate. Therefore, the
judgments expressed here are in p a r t subjective, although my
i n t u i t i o n s were c o n s t a n t l y checked agai ns t those o f o t h e r n a t i v e
speakers and a l o t o f use was made o f the m a t e r i a l drawn from the
newspapers, magazines and books mentioned in 1.4. I consider
t h a t t h e r e would be l i t t l e disagreement as to the s ta t us o f items
marked SI and S3. The former are d istin ctly 'learned', of
'k a tha re v ous a' origin, and t h e i r use i s r e s t r i c t e d t o the speech
o f educated speakers. The l a t t e r are l i n k e d to ' p o p u l a r ' origin,
l i t e r a r y use or r e s t r i c t e d to ' r u r a l speech'.

5. This obs er vati on can be associ ated wi th Verschueren's (1 9 81 : 33 6 ’

7) remarks on the f o l l o w i n g example of a taxonomy:

animal bird robi n C a l i f o r n i a robi n

animal bird sparrow f i e l d sparrow

animal bi rd pigeon fancy pigeon

SUPERORDINATE BASIC LEVEL SUBORDINATE

- 263 -
 Verschueren observes t h a t the

" i n h a b i t a n t s o f some p i g e o n - i n f e s t e d European c i t i e s may be

expected t o have PIGEON as a basic l e v e l term f o r pigeons
and BIRD f o r a l l ot he r b i r d s " .

Verschueren intends t h i s as an example o f the f a c t t h a t ther e is

no simple correspondance between cognitive structure and
biological hierarchy. There i s , however, an a l t e r n a t i v e way of
interpretin g it. For city d w el l e rs in general, ' p i ge o n' is
like ly to be the most characteristic bird they know, i.e. the
most prototypical member of the category 'b ird '. It may
t h e r e f o r e a cqui r e 'basic level' status, i.e. g e n e r a l i z e through
moving one l e v e l up.

6. The noti on of incorporation of abstract, formal components is

a l i e n to the t h e o r e t i c a l framework adopted in the present study.
Ne ver the le ss , the semantic i n fo r ma t io n associ ated with
d irectional/locative Advs can be 'translated' into descriptive
and 'inform al' appropri ateness c ondi ti ons in a fa irly
s t r a i g h t f o r w a r d manner.

7. This is the most conscise presentation of material and

conclusions also included in ot he r works of Rosch, e.g. Rosch
(1976) r e p r i n t e d i n Johnson-Laird ( 1 9 7 7 ) .

- 264 -
 5. ELICITING INFORMATION FROM NATIVE SPEAKERS

5.1 Semantic s i m i l a r i t y s o r t i n g tasks and c l u s t e r a n a l y s i s

As a l re a dy demonstrated, the semantic i nf or ma ti on p e r t a i n i n g t o the

verbs under c o n s i d e r a t i o n i s both complex and o f d i f f e r e n t types. No
a n al y s i s or classification can encompass it a ll, no matter how
detailed it may be. It is also suggested that the overall
o r g a n i z a t i o n o f the domain is complex and t h a t s p e c i f i c p a tt e r n s can
be shown to depend on the items chosen. By i m p l i c a t i o n , any t e s t s
which might be set up to e lic it i nf or ma t io n on structuring and
properties from native speakers w ill be of lim ited use. The
necessity, however, o f o bt a i n i n g such i nf or ma ti on has been stressed
a t many poi nt s throughout the preceding di scussion. Besides, the re
is a lot of evidence from a number of studies, that part of the
i n for ma ti on in the subjective l ex ic on can be r e ve a l e d through
semantic sim ilarity sorting tasks and p r o t o t y p i c a l i t y te s ts .* Both
types o f t e s t s are t h e r e f o r e used and t h e i r r e s u l t s w i l l be r eported
and discussed in what f o l l o w s .

In semantic sim ilarity sorting tasks (hencef or th SST) subj ects are
asked t o group t o g e t h e r items which they consider t o be more c l o s e l y
r e l a t e d i n meaning than they are wi th ot he r items o f the set o f f e r e d
to them. They are t h e r e f o r e r e q u i re d t o understand the meanings of
these items in whatever way available to them and to weight and
combine such c r i t e r i a , i.e. to decide on t h e i r r e l a t i v e importance,
in whatever way they deem fit. The general idea is that such
i n for ma ti on cannot be e x t r a c t e d d i r e c t l y and t h a t i n f e r en c es can be
made on the basis of a careful c o n si d e r at i o n of the resulting
clusters of i tems. One of the methods of processing the data
obtained from individual subjects' groupings has been extensively
used by M i l l e r ( 1969, 1971, 1972), Fillenbaum and Rapoport ( 1 97 1) ,
Long (1975) and others. It involves tabulating co-occurrences of
items in the form o f a m a t r i x and then appl yi ng c l u s t e r a n al y s i s to
this data m atrix, in order to t ransfor m it into clusters. This

- 265 -
method o f o bt a i n i n g a h i e r a r c h i c a l c l u s t e r i n g scheme ( h e n c ef o rt h HCS)
i s the one a p pl ie d here.

A number o f poi nt s have t o be made at the o ut s et concerning the goal

and the appr opr iat eness o f the method used. M iller's original idea
was that clusters obtained by the procedure just mentioned would
somehow correspond to the taxonomic structures proposed by various
t h e o r i s t s f o r the same or s i m i l a r types of items and t h a t t hey would
also reflect t he semantic f e a tur es/ di men si ons involved. This
suggests t h a t i f a s p e c i f i c set o f items involves f e a t u r e s which are
not h i e r a r c h i c a l l y ordered (t a xo no mi ca l ly organi zed) c l u s t e r a n al y s i s
( he nc ef or th CA) would be inappropriate. A further assumption is
considered to underl y t h i s method, namely t h a t items which have been
judged as s i m i l a r (and t h e r e f o r e grouped t o g e t h e r ) have been grouped
on the same b a si s , i.e. t h a t t he same semantic p r o p e r t y was somehow
fe lt as being r e spons ibl e f o r the groupings o f a l l the subj ects who
put the same items t o g e t h e r . The question i s t h e r e f o r e r a i s e d as to
whether such a procedure i s a p p l i c a b l e t o sets of items which are not
necessarily considered to i n vol ve independent and hierarchically
ordered properties and especially to sets of items which may be
judged as s i m i l a r on a number o f d i f f e r e n t semantic c r i t e r i a .

There are two po s si b le answers t o these problems, a theoretical one

concerning the 'n a tu re ' of cluster a nal ys is as a sta tistic a l
technique and an e m p ir i ca l one based on the possi bl e i n t e r p r e t a t i o n s
of the results of specific sets submitted to this method of
processing.

It has t o be noted t h a t CA i s g e n e r a l l y not i nvolved w i t h hypothesis

testing (Anderberg 1 9 7 3 : 1 1 ) . I t can be used as a di scov er y procedure
and can help avoid f o r c i n g or imposing a p a r t i c u l a r s t r u c t u r e on the
data. This suggests, that if for i n st a n c e , the prevailing factor
r espons ib le for groupings is not a p r o pe r ty hypothesized as
hierarchically hi gher than o t h e r s , more than two main c l u s t e r s w i l l
appear on the f i n a l HCS and the hypothesis i n question w i l l be e i t h e r
partly or w hol ly di sca rd ed. Thi s does not , however, imply t h a t we
end up w it h uninterpretable and useless clusters. It simply makes

- 266 -
more d iffic u lt the task of interpreting the results, or rather,
looking for an a l t e r n a t i v e e x pl a n a t i o n to hierarchical or de ri ng of
properties.

To be more s p e c i f i c , c onsi der the example o f ki nshi p terms, which do

not i n v ol ve a strict hierarchy of properties. Since some subj ects
may put 'daughter' and 'son' t og e t h e r on the basis of some
'generation' property, other s on the basis of 'lin e a lity '
(consanguinity), and s t i l l others on the basis of both p r o p e r t i e s ,
the final cluster cannot be characterized by onl y one of these
properties, but w i l l have to r e f l e c t both. Moreover, a f o u r t h group
o f subj ects may put 'daughter' and ' s on' in separate c l u s t e r s on the
basis o f the ' f e m a l e ' vs ' mal e' d i s t i n c t i o n . The f i n a l HCS w i l l show
at least whether the 'sex' pr op e rt y is more salient than the
combination o f ' c o n s a n g u i n i t y ' and ' g e n e r a t i o n ' or v i c e versa (which
w ill depend directly on how c l o s e l y related ' son' and ' d a ug ht e r '
appear t o be, i.e. on t h e i r number o f c o- oc cur r enc es ). I cannot see
why such a f i n d i n g i s l es s i mportant than ' r e l a t i v e h i e r a r c h y ' and I
do not think that it involves a greater idealization than the
'hierarchy t e s tin g ' interpretations.

A number of studi es report the results of cluster-analysed SSTs

c a r r i e d out on sets o f items which do not n e c e s s a r i l y i n vol ve e i t h e r
hiera rchical or de r in g or independence of features (e.g. Fillenbaum
and Rapoport 1971 - verbs of 'judging', Long 1975 - motion v e r b s ) .
These r e s u l t s are both 'plausible' and '1 i n g u i s t i c a l l y relevant' in
t he sense o f M i l l e r (1959:181-2). What i s more i mpor tant, they are
also corrobor at ed by o t h e r methods which do not i n v ol v e h i e r a r c h i c a l
o r d e r in g , such as m u l t i - d i m e n s i o n a l s c a li n g ( Fi l lenbaum and Rapoport,
ib id .).

A final note is in or de r a t t h i s po in t on what we expect the c l u s t e r -

analysed SST t o do in this particular investigation. The overall
o r g a n i z a t i o n o f the domain o f MGMVs i s c l e a r l y d i f f e r e n t l y s t r u c t u r e d
from t h a t o f any s p e c i f i c subset which can be given to subj ects t o
s o r t out. I t i s not t h e r e f o r e pos si b le to expect any t e s t based on a
subset t o reveal the o v e r a l l s t r u c t u r i n g ( or prove the v a l i d i t y o f an

- 267 -
al re ady e s t a b l i s h e d one). It seems, however, pos si b le to check the
r elevance of some properties arrived at through traditional
linguistic analyses a lr e ad y discussed in previous c h ap t er s . It can
be shown, a l so , that s ubj ects can sometimes express the c r i t e r i a on
t he basis o f which they have put items t o g e t h e r . It is interesting
to notice, that such criteria rarely correspond to a single,
independent property. The b elief expressed at many p oi n t s in the
course of the preceding a n al y s i s that certain properties, or
combinations o f p r o p e r t i e s are more s a l i e n t than others can a t l e a s t
be p a r t l y checked by consi de ri ng both the r e s u l t s o f the f i n a l HCS
and i n d i v i d u a l p a i r i n g s o f items (appearing in the form o f the number
of co-occurrences in the data m a t r i x ) .

5.1.1 Semantic s i m i l a r i t y t e s t s i n v ol v i ng English motion verbs

M iller (1972) and Long (1975) present results of SSTs i n vo l v i n g

English MVs. Their results are therefore worth l ooking at more
c a r e f u l l y than those o f o t h e r t e s t s . The general p i c t u r e seems to be
that particularly strong hypotheses were tested and that these
hypotheses cannot be r e f u t e d on the basis o f the results obta ine d.
The p r a c t i c e f ol l ow ed i n both works o f o v e r g e n e r a l i z i n g the v a l i d i t y
of findings based on a particularly small amount of data seems
unwarranted, especially as it is also open to alternative
interpretations. However, two issues r ai se d in M i l l e r (1972) are of
special interest f o r our present purposes, namely r e l a t i v e sal i ence
and the i n c o r p o r a t i o n o f d i r e c t i o n a l components.

One o f M i l l e r ' s SSTs i nv ol ve s 18 ' d i r e c t i o n a l ' English MVs which are

given t o 52 s ubj ects to sort i n t o groups w i t h i n sentence frames of
the form "He V e r b ( X ) " . These sentence frames d i f f e r e n t i a t e between
c aus at ive and non-c aus ati ve uses of the verbs in quest ion. M iller
is interested in demonstrating the psychological reality of the
formal directional components he considers to be i nc or po r at e d in the
verbs, e.g . 'around', 'to ', 'o u t', 'down'. He also examines the
hierarchical r e l a t i o n s between d i r e c t i o n a l p r o p e r t i e s and c a u s a t i v i t y
and concludes that the causative/non-causative distinction
('O b jective'/'R eflexive' in hi s t ermi nol ogy) i s s tr onge st in the ' t o '

- 268 -
cluster, but that the directional component is in general more
salient and hierarchically hi ghe r than t he causative/non-causative
one.

A close l ook a t the actual co-occurrences o f verbs M i l l e r gets shows

th a t such conclusions are not whol ly sust ai ned. Consider f i r s t the
data from the 'downward' motion group o f items, which i nvolves both
c au s at iv e and non-c aus ati ve verbs and is probably the c l e a r e s t case
within this test. Each p a i r o f verbs is fol l owed by the number o f
co-occurrences it has in the m a t r i x , i.e. the number o f subj ects who
put the two verbs t o g e t h e r :

(a ) l owerCAUS - dropCAUS -> 30

descendNON-CAUS " s i n k N0N-CAUS _> 31

f a l 1 N0N-CAUS “ s i n k NON-CAUS _> 41

( b ) d r o p CAUs - f a l l N O N - C A U S ^ 19

l o w e r CAUS - s i n k N 0N-CAUS - > 18

l o w e r CAUs - f a ^ N O N - C A U S _> 15

d r °PCAUS - s i n k N0N-CAUS > 16

(c) d e s c e n d N0N_cAUS " l o w e r CAUS “> 28

descendN0N_CAUS - f a l l N0 N - C A U $ "> 28

The data in (a) and (b) suggest that when both items within the
'downward' group are CAUS or NON-CAUS they get the hi ghe st r a t e s ; if
they do not share the p ro pe r ty they get much l ower scores. In f a c t ,
consi de ra bl y fewer than 50% o f the t o t a l number o f subj ec ts put the
items in question together. The data in (c), however, seem to
suggest that the causative/non-causative distinction does not pl ay
any r o l e .

- 269 -
It must be pointed out that M iller's test does not i n vo lv e any
'upward' motion verbs. The ref or e the p o s s i b i l i t y o f c l u s t e r i n g the
above subset ('downward' motion) w it h any ot he r verb of the set
o f f e r e d to subjects i s r a t h e r low. In the absence o f any 'upward'
direction verbs and on the basis of such l i m i t e d and i nc oncl us ive
evidence, M i l l e r ' s c l ai m t h a t d i r e c t i o n a l i t y is more s a l i e n t than the
c a u s a t i v e / n o n - c a u s a t i v e d i s t i n c t i o n is not wel l founded.

The data from the remaining d i r e c t i o n a l verbs o f t he set are not any
better. The 'around' group consi sts o f what seems to me to be an
i n c l u s i v e term, i.e. 'tu rn ' and i t s hyponyms. C l u s t e r i n g these items
t o ge t h er may be i n t e r p r e t e d as r e v e a l i n g the psychol ogi cal validity
o f the ' t y p e o f ' lexical relationship, i.e. hyponymic i n c l u s i o n .

Wi thi n the group o f verbs c h a r a c t e r i z e d by M i l l e r as incorporating

'out', 'ejec t' and ' wi th dr a w' have only 17 co-occurrences (again
c onsi der ably fewer than 50% o f the subj ects put them together),
although they are both caus at ive s (in M iller'ste s t). On t he other
hand, 'e x it' and 'le av e ', both non-causati ves but belonging to
different 'd irectio n al' groups, namely ' o u t ' and 'away' r e s p e c t i v e l y
(according to M i l l e r ' s c l a s s i f i c a t i o n ) score 42, i.e. hi gher than the
best case w i t h i n the 'downward' group.

Consider, further, the set of verbs M iller classifies under

'to g eth er':

assem blers - c o l l e c t r s -> 38

assembleCAUS - g a t h e r CAUS -> 40

g a t h e r CAUS - c o l l e c t CAUS - > 50

M i l l e r takes these high r a t i n g s as pr oo f o f the e xi s t e n c e (actually

the importance) of t he 'd irectio n al' (as he c a l l s it) preposition
'to gether'. But to start with, these verbs are near-synonyms, so
t h e i r being c l u s t e r e d t o g e t h e r i s not much o f a pr oof f o r anything as
concrete as a 'directional preposition' (although their common
pr ope rty is certainly linked to some c o n di t i o n of 'putting thi ngs

- 270 -
together' and we do not really need a t e s t to prove it). What is
probably more interesting, though, is that although they are all
three supplemented wi th an "X" (in the sentence f ra mes ), the last
pair gets a much hi ghe r score. One possi bl e e xp l an a ti o n could be
t h a t 'assemble' may be associ ated wi th a d i f f e r e n t ( l i n g u i s t i c ) frame
than the ot he r two ( e . g . assemble pieces t o c onst r uct a u n i t , w h i le
'gather' and ' c o l l e c t ' may not be so understood). Be t h a t as i t may,
M iller's second attempt at establishing the 'incorporatio n' of
s p e c i f i c components/features and a f e a t u r e h i e ra r ch y on the basis of
such r e s u l t s seems r a t h e r u n s a t i s f a c t o r y .

Once again the hi ghe st scores correspond to a combination of two

common properties: presence of causativity and 'putting things
together', where t h i s latter pr ope rt y is certainly not o f the same
s tatus as 'downward' or ' a r o u n d ' , i.e. pu re l y ' d i r e c t i o n a l ' .

M iller also considers the 'to g eth er' group in connection wi th

'approach' and 'v is it', the former of which he classifies under
' t o wa r d' and the l a t t e r one under ' t o ' . Notice now t he score o f t h i s
l a s t p a i r : approach X - v i s i t X -> 26 and M i l l e r ' s comment:

"The r e f l e x i v e - o b j e c t i v e d i f f e r e n t i a t i o n i s str onge st f o r the t o

clu ste rs: a p p r o a c h e s and v i s i t s were c l u s t e r e d t og e t h e r wi th
a s s e m b l e s , c o l l e c t s and g a t h e r s by only one or two judges. Since
the d i r e c t i o n a l p r e p o s i t i o n s are r e a l l y tow ard and t o for
a p p r o a c h e s and v i s i t s (hence a c e r t a i n r e lu c t a n c e to put them in
the same p i l e ) and t o g e t h e r f o r the o b j e c t i v e motion verbs
a s s e m b l e s - c o l l e c t s - g a t h e r s , the tendency to keep them a pa rt may
r e f l e c t the d i r e c t i o n a l component as wel l as the r e f l e x i v e -
o b j e c t i v e ( or c a us a t i ve ) component", ( p . 358)

But the score o f the 'approach' - 'v is it' p a i r i s almost the average
of what he gets in the particularly strong 'downward' group and
corresponds exactly to 50% o f the answers, although according to
M iller's analysis, 'approach' and 'v is it' are nowhere near as close
in ' d i r e c t i o n a l i t y ' as the groups e x h i b i t i n g hyponymic r e l a t i o n s , the
'downward' motion group and the near-synonyms a lr e ad y discussed. As
for the 'together' group, it is doubtful whether it contai ns verbs
mainly understood as motion verbs and equally doubtful whether

- 271 -
'together' is to be understood as synchronicallyre la te d to the
pur el y ' d i r e c t i o n a l ' 'to '.

Consider fin a lly the 'rotary motion' set results which g e t a great
number o f co-occurrences, i n d i c a t i v e o f the r e l a t i v e salience ( i n my
opi ni on) of the p r o pe r t y of 'rotation' or 'change of orientation'
compared t o o th e r 'd irectio n al' properties (such as those s i g n a l l e d
by ' t o ' , ' a w ay ' , e t c . ) :

(a) t u r n N0N_CAUS - p i v o t N0N. CAUS -> 48

r o t a t e CAUS - s pi nCAUS -> 47

(b) p i v o t N0N. CAUS - s pi nCAUS - > 30

r o t a t e ^ u s - PivotNON-CAUS _> 29

t u r n NON-CAUS ' ™ t a t e CAUS -> 29

t u r n NON-CAUS ' sPi n CAUS 28

It is again f a i r l y c l e a r t h a t the c a u s a t i v e / n o n - c a u s a t i v e d i s t i n c t i o n
plays a d e c i s i v e r o l e , otherwi se the d i f f e r e n c e s o f the scores in (a)
and (b) are not e x p l a i n a b l e .

To sum up, M iller's c lai m that " the directional component is far
more s a l i e n t than t he reflexive-objective one and might be said to
dominate it hierarchically" (M ille r 1969:176, 1972:358) is not well
sustained by t he evidence he provides, through t h i s test, especially
if one takes a close look at specific co-occurrences. Ce r t a i n
properties linked with ' d i r e c t i o n a l i t y ' are more s a l i e n t than o the rs .
Combinations of causativity and d irectionality are clearly most
salient. General conclusions pertaining to relative salience and
especially to r e l a t i v e hierarchical order are d i f f i c u l t t o draw on
the basis o f a r a t h e r small pi ece o f d a t a . As a lr ea dy pointed out,
the o r g a n i z a t i o n o f the domain as a whole w i l l be d i f f e r e n t from t h a t
of any s p e c i f i c subset; and a h i er a r c h y o f p r o p e r t i e s (if distinct
from s a l i e n c e ) could be b e t t e r approached in an i n v e s t i g a t i o n of a

- 272 -
much larger cont ext (a whole domain, rather than any specific
subset).

If one wishes t o check the psychol ogi cal validity o f a major (high
level) pr o pe r ty such as c a u s a t i v i t y , for i n st an ce , one i s obl iged to
provi de a fa irly large number o f items to be sor ted o ut . This is
p r e c i s e l y what M i l l e r (1969, 1971) does in the case o f nouns, wishing
to test the validity of the 'object' - 'non object' distinction.
Subjects are given 42 items to s o r t out. M iller's observati ons on
the r e s u l t i n g c l u s t e r s are very eloquent:

"Did the semantic marker [ ' o b j e c t ' , 'non o b j e c t ' ] t h a t was

d e l i b e r a t e l y i ntroduced i n t o the set o f words reappear in the
a nal ysis? Yes and no. The c l u s t e r s obtained d i d not c o n t r a d i c t
the hypothesis t h a t our judges were s o r t i n g w i t h t h i s semantic
d i s t i n c t i o n i n mind, y e t t h e i r data i n d i c a t e a f i n e r a nal ysis
i n t o a t l e a s t f i v e , r a t h e r than only two c l u s t e r s , so the obj ect
marker i s not compl etel y v e r i f i e d by these d a t a " .
( M i l l e r 1971:577)

In practical terms this means that subj ects tend to form small
groups, e s p e c i a l l y when faced w i t h more than 10 or 20 items which are
not governed by obvious hyponymic r e l a t i o n s h i p s , and t h a t therefore
high l e v e l distinctions are somehow l o s t . This does not mean t h a t
they have no psychol ogi cal reality or that they are not
hierarchically hi ghe r than ot he r distinctions or properties. It
probably means t h a t h i e r a r c h y cannot be t e s t e d in t h i s way, but t h a t
relative salience of combinations of properties can. Smaller
clu sters and s u bcl us te rs provi de fa irly reliable evidence to this
effect.

An i mportant issue f o r the present study is raised i n Long ( 1 9 75 ) ,

namely the relation between h i e r a r c h i c a l or der (of properties) and
number o f shared p r o p e r t i e s . An obvious question i s which p r o p e r t i e s
are taken into account ( si nce they are not directly provided by
s u b j ec t s ) and a related one is whether they correspond to a most
general or a particular understanding of the items tested. Long
makes the assumption that features (properties) are of equal
importance, which is not accepted here and which she recognizes
h e r s e l f to be an o v e r s i m p l i f i c a t i o n . A careful look at her r e s u l t s

- 273 -
shows t h a t re la tiv e salience of certain combinations of properties
provides a better e x p la n at io n of the data than number of shared
attributes and t h a t , as in M iller's tests, specific co-occurrences
are worth consi der ing in d e t a i l (besides the o v e r a l l HCS).

Long's study i nv ol ve s 15 English MVs analysed on the basis of

M iller's (1972) classification but s l i g h t l y modi fi ed and r e s t r i c t e d
to those specific meanings of the verbs which corresponded to the
a ct ions shown t o subj ec ts on a videot ape . Two groups o f subjects are
asked to sort the verbs out. One group has seen the videotaped
actions, the other one performs a purely verbal test, i.e. a SST
based on the same set o f verbs w r i t t e n on separate cards w i t h ou t any
supplementary i n fo r m a t i o n as to how they are t o be understood. The
results are shown on a s i m i l a r i t y matrix ( p r e se nt in g the number of
co-occurrences for each p a i r o f items) which i s f u r t h e r transformed
i n t o c l u s t e r s by appl yi ng CA. The s i m i l a r i t y m a t r i x i s compared to a
matrix o f shared features (for each p a i r o f v e r b s ) , which Long has
const ruct ed on t he basi s of the a n al y s i s a lr e a d y mentioned (a
modification of M i l l e r ' s 1972 components and the act ion s presented on
the v i d e o t a p e ) .

Despit e the fact that Long's hypothesis is a ls o very strong

(correlation o f shared a t t r i b u t e s to f e a t u r e - h i e r a r c h y ) and her own
remark t h a t i t i s not a p p r o p r i a t e t o t e s t the c o r r e l a t i o n between the
two measures s ta tis tic a lly "since the numbers within each of the
mat ri ce s are not c ompl etel y independent" ( i b i d * :58), the c o r r e l a t i o n
is so high (.80) that Long's hypothesis cannot be r e j e c t e d . This
means t h a t the possibility of c orrela ting the number o f attributes
a r r i v e d a t on the basis o f a t r a d i t i o n a l l i n g u i s t i c a n a l y s i s wi th the
number o f co-occurrences in an SST m a t r i x cannot be r ul ed out . The
correlations that are o f i n t e r e s t here concern the judgments o f the
20 subj ects performing the verbal SST (what Long c a l l s t he 'si mple
sort group'). The maximum number o f co-occurrences i n the SST m at ri x
would be 20 (i.e . 20 subjects). The maximum number of shared
a t t r i b u t e s / p r o p e r t i e s in Long's ma tr i x is 4.

- 274 -
Each p a i r is f ol l ow ed by the number o f co-occurrences i t gets in the
SST matrix and then by the number of shared properties it has
according to Long's a n a l y s i s . The best c o r r e l a t i o n s she gets are:

jump - hop -> 18 : 4 ( - L o c, - D i r . , Manner, s p e c i f i c Manner)

push - p u l l -> 17 : 3 ( S p e c , d i r e c t . , + D i r . , -Propellent)

Notice, however, that Long h e r s e l f observes ( i b i d . :103) that a pair

o f words may d i f f e r along several dimensions and o f f e r s jump-hop as a
case in point, suggesting t h a t they d iffer in the number of feet
used, in 'd irectio n ' ( ' u p and down' for 'jump' vs ' a l o n g ' f o r hop),
etc. So even t he p a i r which gets the hi ghest number of co­
occurrences (and the best correlation with number of shared
attributes) is considered as a possi bl e example of 'multiple
contrasts'. Compare the co-occurrences o f both items t o ' d a nc e' :

jump - dance -> 14 : 3 (Manner, -D ir., -Loc.)

hop - dance -> 13 : 3 (Manner, -D ir., -Loc.)

It can be argued t h a t , u n l i k e the remaining items o f the set o f f e r e d

to s u bj e ct s , a ll these three verbs i nv ol ve a pr op e rt y o f 'relative
absence o f seriousness o f purpose' (as suggested in 4 . 2 . 4 ) . This can
e xpl ai n why over 50% o f the subj ects c l u s t e r them t o g e t h e r . Besides,
although 'jump' and ' h o p ' , may be said to d i f f e r on a number of
points, they do refer to particularly sim ilar motions, especially
compared wi th the remaining verbs of the set. This p r op e rt y ,
combined w i t h t he one a lr ea dy mentioned i s probably enough to keep
them t o g e t h e r as the str onge st p a i r in t h i s t e s t . The e xpl ana ti on o f
'number o f shared attributes' cannot be considered separately from
'type of a t t r i b u t e s ' and t h e i r r e l a t i v e s a l i e n c e .

To a ppr ec i at e this, compare the items a lr ea dy discussed wi th the

'push-pull' pair, which is also considered to i n v ol v e 3 shared
properties but has almost as many co-occurrences as the 'jump-hop'
one ( s ai d to involve 4). Antonymous pa i rs can be accounted f o r on

- 275 -
 the basis of 'one feature difference' but especially if only 4
features (properties) are taken into account, t he difference in
question becomes d i s p r o p o r t i o n a t e l y g r e a t . There i s strong evidence
( t o be discussed in the SST on MGMVs) t h a t antonyms or near-antonyms
are only separated if the possibility arises of i n cl ud i ng either
member in a group marked f o r a particularly salient pr ope rt y (e.g.
downward/upward m o ti o n ). Number o f shared p r o p e r t i e s cannot e x pl a i n
t h i s tendency o f s u b j e c t s ' groupings.

Long's worst c o r r e l a t i o n s are also worth c onsi der ing c a r e f u l l y . In

order to a p pr e c i a t e some o f these, it should be noti ced that both
groups o f subj ects ( t h e ' s im pl e sor t group' and the 'taped s o r t ' one)
produced s i m i l a r results, wi th the notabl e exceptions o f 's pin' and
shake . To s t a r t w i t h , this i mpl ie s t h a t subjects perform the t as k
w i t h p a r t i c u l a r and not general meanings in mind (Long 1 9 7 5: 6 3) . The
videotaped act ions corresponding to the items t e st e d are p a r t i c u l a r l y
'good i nst an ce s' of the c a t eg or i e s in quest ion, as can be e a s i l y
attested in Long's description of these act ions (ib id .:3 5). This
constitutes in its e lf very good evidence in favour of the
prototypical rather than the most general understanding of items,
although Long is not concerned w it h Prototype t he or y.

The two exceptions are p r e c i s e l y ' s p i n ' and ' s h a k e ' , where the former
a c t io n i nvolved stepping in c i r c l e s in one pl ace and the l a t t e r one
standing in one p l a ce , holding a b a l l out in f r o n t o f the body wi th
both hands and shaking i t v i g o r ou sl y up and down. Long's a n al y s i s o f
the corresponding verbs in the form o f f e a t u r e s i s based on the above
understandings of t hese verbs. Ther ef or e the fact that onl y these
items in the set offered can be understood e i t h e r as causatives or
non-causati ves cannot be taken into account. The ambiguity a r i s es
only for the ' s i mp l e sort group' who have not seen the videotaped
actions.

In t h i s l i g h t , consi der b r i e f l y some o f Long's worst c o r r e l a t i o n s :

- 276 -
 shake - hit -> 11 : 4 h i t - carry - > 9 : 3

shake - spin -> 11 : 0 c a r r y - drop -> 14 : 1

shake - c a r r y - > 9 : 3 drop - k i c k - > 6 : 2

The verbs 's ha ke ' and 's p in ' are completely isolated in Long's own
hierarchical a n a l y s i s , which draws a l i n e between caus at ive s and non­
causati ves and never considers intra-group shared p r o p e r t i e s , hence
the 0 number of shared properties, which predictably does not
correlate with the number of co-occurrences (about 50%). On the
o th e r hand 'shake' and 'h it' are attributed the maximum o f common
properties, as the non-causati ve use of ' shake' is ignored.
Something s i m i l a r a p p l i e s to the 'shake' - 'c a rry ' pair.

The verbs 'h it' and 'carry' are in tuitively fe lt to be rather

dissim ilar. Even the shared element o f c a u s a t i v i t y i s less c l e a r in
the case o f ' h i t ' (see 3 . 1 ) . Only the l a t t e r verb i s a c l e a r case o f
a 'causative of motion'. If 'h it' is also considered a caus ati ve o f
motion, then '-P ropellent' is erroneousl y included by Long as a
common p r o pe r ty o f the pair. On the ot he r hand, t he s im ilarity of
'carry' and ' dr o p ' can be most e a s i l y expl ai ned in terms o f frames,
in the sense of Brown and Yule ( 1 9 8 3 :2 3 8- 4 7) or Verschueren' s
(1981:338) 'prototypical scenes' / ' frames' . In common terms one can
drop something t h a t one i s c a r r y i n g . If number o f a t t r i b u t e s is t o
be used as a measure of semantic sim ila rity, then '-Propellent'
should be a common pr op e rt y in addition to 'non-causative'. The
prototypical understanding of 'drop' is more likely than not be
c h a r a c t e r i z e d by t h i s s a l i e n t pr ope rt y ( c o n t r a r y to Long's a n a l y s i s ) .
This could also provi de a b e t t e r c o r r e l a t i o n in the case o f the l a s t
pair 'dro p-kick', which would be shown as sharing onl y the caus at ive
element and i n d i f f e r e n t degrees a t t h a t .

To sum up, Long's study provides f a i r l y reliable evidence i n favour

of a prototypical r a t h e r than a general understanding o f c at e go r ie s
(and o f those p e r t a i n i n g t o MVs in p a r t i c u l a r ) . It also raises an
i n t r i g u i n g question as t o the r e l a t i o n o f number o f shared a t t r i b u t e s
resulting from a t r a d i t i o n a l linguistic analysis o f MVs t o semantic

- 277 -
sim ilarity judgments of native speakers. It seems reasonable to
expect t h a t such c o r r e l a t i o n s are not compl etel y i mpos si bl e, provided
t he 'rig h t' s o rt of attributes are taken into account (i.e.
prototypical rather than general ones) and the issue of relative
s al i e n ce i s allowed t o pl ay a d e c i s i v e r o l e .

5.1.2 A semantic s i m i l a r i t y t e s t wi th 34 Modern Greek motion verbs

As a l re ady poi nted out in 5.1.1, particularly strong hypotheses

cannot be e a s i l y tested wi th cluster-analysed SSTs. On the o the r
hand, if quite a l o t i s known about the items t e s t e d , CA need not be
used merely as an exploratory devi ce , A mi ddl e- course seems
therefore most appropriate. We cannot expect to obt ai n results
d e c i s i v e f o r the o r g a n i z a t i o n o f the whole domain but can i ncl ude a
fa irly l a r g e number o f items which we consi der r e p r e s e n t a t i v e o f the
overall domain. The i nf er enc es t h a t can be made may i n v ol v e number
of a ttr i b u te s , but these a t t r i b u t e s need not be e i t h e r homogeneous or
independent or of equal wei ght . The SST to be descri bed in what
f ol l ow s was conducted w i t h these p r e l i m i n a r i e s in mind, some o f which
are the d i r e c t outcome o f the preceding obser vati ons on M i l l e r ' s and
Long's work.

Some o f the p r o p e r t i e s discussed i n 4 .2 are a p p l i c a b l e both to a most

general and a p r o t o t y p i c a l understanding o f the r e l e v a n t ver bs. At a
number of poi nt s differences between the two understandings are
pointed out. Checking all these properties would be an impossible
task for any s i n g l e test. It seems, however, feasible to attempt
checking the validity o f what have been posi ted as l a r g e distinct
c at e g o r i e s w i t h i n t h i s area and t h e i r i n t e r r e l a t i o n s . These are:

- Causatives vs non- ca us at iv es, agentives vs non-agentives

- C ha n g e - o f - l o c a t i o n vs c h a n g e - o f - p o s i t i o n and
' d i s t a n c e c o ve r i n g' w i t h i n the CL group

- 278 -
 - General motion vs 'manner' and whether a d i f f e r e n c e in
'medium' plays an impor tant r o l e or not

- Vertical vs i nd e t e rm i n a t e d i r e c t i o n , change o f o r i e n t a t i o n
vs 'random w a l k ' , 'dependent m o t i o n ' , ' p a t h ' ( passage).

a. Choice o f data u n i t s

Fifty subj ects were chosen, all o f whom were n a t i v e speakers o f MG

residing i n Athens (but not n e c e s s a r i l y brought up i n Athens), aged
between 20 and 50. P i l o t work p r e v i o u s l y conducted had shown t h a t a
certain standard o f education was a b s o l u t e l y i n di spe ns ab le : subjects
should be able to classify on the basis of specific instructions
(listed below) and understand, for i n s t an c e , the d i f f e r e n c e between
'semantic sim ilarity' and f r e e - a s s o c i a t i o n (once it is e xpl ai ned to
them). The onl y 'c a teg orie s ' o f people excluded were l i n g u i s t s and
philologists (as they could have had preconceived ideas on the
material, or ' s p e c i a l i s t ' knowledge). Since t h er e was no p a r t i c u l a r
scheme of sampling, the sample can be considered random and
r e p r e s e n t a t i v e o f the educated p a r t o f the p op u l a t i o n .

b. Choice o f v a r i a b l e s

The f i f t y subj ects were each given 34 randomly arranged cards w i t h a

MGMV w r i t t e n on each car d. The v a r i a b l e s (terms used) were r a t h e r
carefully chosen so that they would 'conform' to the following
principles: (a) They should be very common MVs in everyday use
p r ese nt ing 'basic level' categories. (Hyponymic relations are
checked in the test on ' P r o t o t y p i c a l i t y ' in a subsequent section.)
(b) They should be q u i t e representative o f the c a t eg or i e s discussed
in 4.2 (also appearing in List V III) and supplemented w i t h some CL
caus at ive s of motion. Two c a t eg or i e s were not r epr esented as the
m a t e r i a l t hey contained seemed t o me l es s common in everyday use than
the remaining items. These appear in L i s t V I I as ' P a r t i a l motion not
resulting in different p os t u r e ' and 'P artial motion (which is)
regular/repeated'. Only two genuine synonyms were i ncluded:
' p e r i f e r o m e ' and ' t r i y i r i z o ' (roam around). The reason f o r i n c ludi ng

- 279 -
these was to check whether morphological distinctions (e.g. '-o ' vs
' - ome ' endings corresponding normally to active vs mediopassive
forms) would affect the categorization (i.e . subjects' j udgments).
The pair ' anevazo' - ' ipsono' receive identical translations in
English (i.e . raise) but are not genuine synonyms. They were both
included in or de r to check whether phonological f a c t o r s would a f f e c t
subjects' judgments (e.g. make subjects group ' a n e v a z o ' wi th
'a n e v e n o ' (ascend) rather than 'ip s o n o '). Two non-causati ve
transitives were also i ncluded: 'p e rn a o ' (pass) and 'h ia s x iz o '

(traverse) to check whether subj ects would classify them wi th

c a us at ive transitives or non-causati ve in transitives. A number of
p h on ol o g i c a l l y identical causative/non-causative pairs was also
included w i t hout any i n d i c a t i o n as t o which one ( i . e . which member of
the pair) subj ects were supposed to take into consideration. The
result is particularly interesting concerning the direction of
l e x i c a l i z a t i o n a nd/ or Prototype t he o ry .

c. Instructions

Subjects were gi ven the usual i n s t r u c t i o n s f o r such tasks which are

t r a n s l a t a b l e i n t o English as f o l l o w s :

1. Put these verbs i n t o groups ('categories') on the basis

o f t h e i r semantic s i m i l a r i t y .

2. Take i n t o account t h e i r l i t e r a l meaning onl y and not any

f i g u r a t i v e uses they might have; a l s o , t h e i r most general
meaning, i . e . do not t h i n k o f a p a r t i c u l a r sentence or
'im age',

3. Each verb can belong t o onl y one group, t he one to which

i t is most a ki n .

4. You can make as many p i l e s (groups) as you wish; a verb

can be the only member o f a cat egory.

5. The time you w i l l need f o r sorting these verbs into

groups i s o f no importance.

- 280 -
P ilot work had also shown that subjects had to be given an
e x p l a n a t i on and an illu stratio n o f the difference between semantic
s im ilarity and free-association (see Rosch's instructions for
prototypicality tests). This was done o rally and they were then
asked if they had any quest ions. Most of them did the task
individually, at different times and in different p l ace s. After
completing t h e i r groups they were asked to comment on the groupings
if they wished and a note was taken o f what they had to say. The
e xpl ana ti ons o f f e r e d f o r some groupings are p a r t i c u l a r l y interesting
and w i l l be discussed s e p a r a t e l y . A common complaint was t h a t i t was
d iffic u lt not to think of any particular meaning, sentence or
'picture' associ ated wi th a given verb.

d. C l u s t e r a n al y si s

A minimum o f 3 and a maximum o f 17 groups were formed; t he mean o f

p i l e s was 9 . 9 2 . A m a t r i x o f s i m i l a r i t y was formed showing f o r each
p a i r o f verbs the number o f subj ects who placed t h a t p a i r i n the same
group. The m a t r i x was then converted i n t o a h i e r a r c h i c a l clu stering
scheme, following Johnson ( 1 9 6 7 ) . The basic p r i n c i p l e o f Johnson's
method is t h a t a t each step, a variable is l i n k e d w i t h the c l o se s t
cluster or (other) variable. Starting wi th 34 variables in this
p a r t i c u l a r case we merge them step by step t o one c l u s t e r . (Both the
m a t r i x and the HCS are reproduced o v e r l e a f . )

At p o i n t 19 o f the HCS, s i x c l u s t e r s can be i d e n t i f i e d , l a b e l l e d here

A - F and i n c lu d i n g the f o l l o w i n g items:

A: items 6 to 8
kaOome, k s a p l o n o , yonatizo, sikonome

B: items 32 t o 30
s ka rfa lo n o , aneveno, kateveno

C: items 5 t o 33
k ila o , y lis tra o , v u l j a z o , sernome

- 281 -
1) k s a p l o n o 0 11 1 38 4 1 1 0 2 1 1 3 1 9 1 1 0 0 2 1 0 1 0 2 1 0 26 0 40 5 2 0 5 3
lie
2) f f l i s t r a o 3 24 8 6 9 7 4 4 7 3 3 6 3 30 3 3 3 0 1 2 0 3 0 6 3 0 2 0 340 5 0 2
slip,slide
3) f e r n o 1 0 0 0 0 0 0 0 9 0 2 0 11 0 0 0 0 1 14 0 11 1 10 1 2 48 0 12 02 1 1

Matrix
bring
4) s t r i v o 9 4 6 1 8 9 10 3 6 10 8 0 15 3 15 15 8 1 0 8 1 8 1 11 8 0 2 0 2 7 22
turn
5) k i l a o 2 22 6 6 10 6 3 3 7 2 4 6 2 26 2 2 2 2 2 1 2 2 2 5 2 2 2 2 31
roll

of
6) kaOome 0 7 1 36 4 1 1 0 2 1 1 2 1 6 1 1 0 0 2 1
1 0 1 0 2 1 0 30 0 3
sit

co-occurrences
7) t i n a z o 0 0 2 0 1 1 •0 21 3 0 2 22 0 2 0 0 0 18 31 0 20 0 17 0 0 13 1
shake up
8) si k o n o me 2 8 13 28 7 3 2 2 6 3 17 0 2 4 2 2 2 10 0 3 3 3 9 4 3 0 2
rise
9) s t e l n o 1 0 0 0 0 0 0 10 0 1 0 11 0 0 0 0 2 10 15 0 11 2 10 1 1 0
send
keno 19 3 7 1 11 11 17 2 6 22 10 0 17 2 12 13 20 0 0 12 0 49 0 11 0
enter
perpatao 21 9 11 2 8 35 14 7 20 16 9 0 14 2 16 16 13 0 0 16 0 11 0 0
walk
anevazo 0 0 13 0 5 0 0 10 2 0 14 20 0 2 0 0 0 48 14 0 31 0 0
take up
vj [eno 19 3 7 1 11 11 17 2 6 22 10 0 17 2 12 13 21 0 0 12 0 0
exit
kr emao 0 0 6 0 3 0 0 81 0 6 20 0 2 0 0 0 30 18 0 10
hang
ak o l u Qo 21 3 4 1 6 10 19 1 4 27 6 0 19 1 19 19 19 0 0 15
follow
rixno 0 0 0 2 2 1 0 22 3 0 0 28 0 4 0 0 0 16 2
throw
i psono 0 0 12 0 4 0 0 9 2 0 13 20 0 2 0 0 0 0
raise
fevjfo 37 4 4 1 7 13 22 3 4 39 6 0 17 2 15 15 1
leave
perfferome 15 6 4 1 5 10 23 2 5 18 5 0 25 3 50 0
roam around
trip riz o 15 6 4 1 5 10 23 2 5 18 5 0 25 3 0
roam around
vuljazo 2 15 5 10 13 3 2 14 2 4 11 2 3
sink
Jiasxizo ‘ 19 4 5 1 S 9 39 2 4 IS 9' 0 4
traverse
gremizo 0 1 4 4 8 0 0 8 0 0 3 6
pull down
ice)
24) aneveno 11 3 •39 2 35 7 7 5 14 7 0
ascend
25) erxotne 38 4 5 1 7 13 21 3 5 1
come
26) p i f a o 4 8 15 3 11 24 5 18 6
jump
27) p e t a o 2 4 7 0 3 10 3 1
fly,throw
28) p e r n a o 19 4 5 1 8 9 2
pass
29) t r e x o 11 10 13 1 9 3
run
30) k a t e v e n o 8 5 23 9 0
descend
31) i o n a t i z o 1 12 1 2
kneel
32) s k a r f a l o n o 4 11 0
cl imb
33) ser nome 4 5
crawl,creep
34) p i t f e n o 0
go
34 33 32 31 30 29 28- 27 26 9 8 7 6 5 4 3 2 1
 xxxxxxxxxxxxx X
t< X
a S3
XX x x x x x x x : ....
..................i x x x x x x : XX
K
XSXSXSXSXS S lS...
iK S l £3
:x x x x x X X X X X X X Xi x x x- x ■ x..x ......x x
.......
....
xxxxxxxxxxxxx X X X X x x x x x x x x ; i x x x x x x x x x x : X X X X X X X X X:x x x x x X X X X X X X Xi x x x x x x x x
XXXXKKXXXXXXX X X X X x x x x x x x x : : X X X x X X X X X X :X X X X XX ;x X X X X X X X X X X X X : x x x x x x x x
x x x x x x x x x x x x x X XXX x x x x x x x : i x x x x x x x x x x :X X X X X X X X X x X X X X X X X X X X i x x x x x x x x
x x x x x x x x x x x x x : x : I X X X X X X X X X X x ; : x x x x x x x x x x ; :x X X X X X X X X X X X X X X X x x x x x x x x
XXXXXXXXXX X X XX X X XX X ;x x x x x x x : x x x x x x x x x x x x x x x x xxxxxxxx X X X X xxxxxxxx
X X X X X X X X X X X X X X x Xx Xx Xx x x x x x x x x x x x x x x x x x x x x x x x x x x x xxxxxxxxxxxxx x
......x
...... ......x
.......X X X X
xxxxxxxxxxxxxxxxx x x : :x x x X X X X X X XX X XX XX X X X X X X X X X :x x x x X X X X
xxxxxxxxxx: x x : :x x x X X X X X X X X XX - X X XX :x x x x X X
xxxxxxxxxx: x x: :x x x X X X X X X X X X X X X X X X X XX X X x x x x :
.xxxxxxxx XXxx : :x x X X X XX ixxxxxxxx X X X X XXXXXXXXX XX xX xX xX : XXX
cxxxxxxxx XXx x x x : :x x x x x x > :xxxxxxxx X X X X X X X X Xx x x :
xxx xx x; ;x x x x x : x x x x x x x x : X XX XX X X X
:x x x x x x x : :x x x x X X X X XX
X X x X X X XXX x x: i xXXXX :x x x x xxxxxxxx: :x x x x :
X X X X X X XXX X X X X X X X X : :x x x x x x x x : ;X X X X X X xxxxxxxx: :x x x x :
X X X X X X XXX X X X X X X x x : :x x x x X X x x : X XX X XX xxxxxxxx;
xxxxxxxxx. ixxxxxxxx: :x x x x X X x x : X X X XX X xxxxxxxx:
xxxxxxxxx: ixxxxxxxxx: x x x x X X X x x x x x X X x X X X X X X X X XXX X XXX
:x x x x x : ixxxxxxxxx
X XX X X XX XX X X X X X XX X x X X
X X X X X X X X X X X X X X X X , . , , — —
x x x x x x x x x x x x x X X XX XX
XXXX
XX X
XXXXXXXXXXXXX
xxxxxxxx XX XXX.
xxxxxxxxxxxxx:
.. I X X X X X X X X X X :
JXXXXXXXXXXX;
XXXXXXXXXXXX x xxx :
X XX X X X
:x x x x x x .
X X X X ; . . . :
.x x x x x x :
IXXXXXXXX
_ XXXX
. X X X X
xxxx
..... KXx>x<X x>x<>x< x x>x
>< <X xX : X X XXXXXXXXXX* I XXXXXXXX X XX X X X
x XXXXX xxxxx
XXXX
Xx Xx Xx X XXXX
xxxxxxxxxxxxxxxxxxxx: ixxxxxxxxxx: X X X X X X X X X X x x x ;
XXX XX X :x x x x x : :x x x x x x ; :x x x x X X X X X X X X XXx :
XX XXX X ;x x x x x x : .............................: x x x x x x X XX Xx X ;x x x x
XXX X X X :x x x x x : :x x x x XXXXXXXXXXXX X Xx x x x x x :
X XX X X X C X X X X
c-t* x x x x x x ^ . ^ . ^ , . X X x x x x > <> 5 X X X X X X i XX XX
xxxxxxxxxxxxxxx: X X . X X x x x x : X XX X X X :x x x x
CO Q .XX X X X X X X X X X X X XX. :x x x x x x : X XX X X X ;x x x x
co "S ;x x x x x x X XX X X X X x x : C X X x x x x : X XXX XX IXXXX
Lx
X X X X X X X XX X X X X x x :
‘ X X X X X ­ I X XX X X X X x x :
IX X X X X X :
;x x X X X X x x :
X X X X XX :x x x x
X X X X XX
x x x x x x ^ - , . ^ : x xx x x x x x x x x x x xx : X X X X XX
xxxxxxxxx xxxxxxxx: xxxxxxxx x x : X X Xx XX X X X X X X
xxxxxxxxx :xxxxxxxxx: xx : X X XX XX X X X X X X
xx ; X X XX X X X X X X X X
:x x x x x :x x x x x x : XX XX xx : X X XX X X
X X X X X x x: :x x x x x xx :
X X X X X x x : xx :
x x x x x x x ; xx : X X X X X X
XX X X X X X XX X X X Xx : X X X X X X X X XX X X
XXX: xx : X X X X X X X X X X X X
xx x : xx ; X X X X X X X X X X X X
xx x ; x x xx x x :x x x x x X X X X xx : X X X X X X XX
X x xx x x xx; X X X X X X XX.
X X xX X X X X xx : X X X X X X X X x X XX
X X XX X X XXX X X X X X X X X x xx ; X X X X XX
X X XX X X XXx X X X X X xx : X X X X XX
X X XX X X XX X X x x xx : X X X X XX
X X XX X X X X X x xx ; X X X X XX
X X XX X X xx : XXXX X X
X X XX X X xx : XXXX X X XXXX
ro X X XX X x XX XX x x : X X X X
co X X XX X X xx : X X X X
X X XX X X xx ; XXXX ., x x xx
XX X X XX XX xx : X X X X XXXX X X X X XX
XX XX xx : XXXX XXXX ' •5X X X X
XX XX xx : XXXX X X X X cxxxx
X X X X X X X X X X X X CX X X X
X X XX X X x X :x
CXxXxXXx
X X X x ;X
CXxXxXXx
X X XX :x x x x
X X X X X X X X :x x :x x x x
X X X x X x X X ;x x ;X Xx x
X X XX X X X X Xx
X X X X X X X X XXXX
X X X X XXXX
xX XX XXXX
X X XXXX
XX X X X X
XX
X X XX

X X Xx

X X X X
. :x x :
X XX x
- ;x x ;
:x x :
XXXX
: x x : -
x x x x
X X X X

X X
X X X X
X X X X
X X X X
X X X X X X
X X X X X X
X X X X X X
X X X X
X X

X X
X X
X X
X X
X X
X X
xrx X X
X X
X X X X
X X

X X
X X
X X

5>
 D: items 16 to 14
rix n o , tin a z o , grem izo, p e ta o , ipsono , a n e v a z o , kremao

E: items 10 to 26
beno, p erife ro m e , triy iriz o , pernao, S ia s x iz o , p iy e n o ,
erxome, fevyo, akoluQo, tre xo , perpatao, piftao

F: items 9 and 3
s te ln o , ferno

The content of these clu sters can be tentatively specified as

follows:

A (6 - 8) c hange-of-pos tur e (CP) non-causati ves

B (32 - 30) v e r t i c a l c h a n g e - o f - l o c a t i o n (CL) non-causati ves
C (5 - 33) non- causati ves, contact of major p a r t of o b j e c t wi th
environment, smooth motion
D (16 - 14) c h a n g e - o f - l o c a t i o n (CL) causati ves
E (10 - 2 6) c h a n g e - o f - l o c a t i o n (CL) n on -c aus at iv es, i nd e te rm in a te
direction
F (9 - 3) d e i c t i c CL c au s at iv e p a i r

One verb ' s t r i v o ' (turn) i s not included in the above c l u s t e r s as i t

joins c l u s t e r s D - F a t a very l a t e stage (15) and c l u s t e r s A - C at
( 8 ).

At poi nt (27) c e r t a i n subcl uster s can be i d e n t i f i e d : D i ncludes two

s u b c lu s te r s, one c o n si s t i n g of items 16-23 'rix n o ' (throw), 'tin a z o '

(shake up), 'g re m iz o ' ( p u l l down) and another one c o n si s t i n g o f items

17-14 'ip s o n o ' (raise), ' anevazo' ( t ak e up), 'kre m a o ' (hang). E
i ncludes t h r e e p a i r s and 2 more s ubcl uster s: items 10 and 13 'b e n o '
(enter), 'v y e n o ' (e x it), items 19 and 20 'p e rife ro m e ' , 'triy iriz o '

(roam around), 28 and 22 ' p e r n a o ' ( pa ss) , 'S ia s x iz o ' (traverse). The
two l a s t p a i r s form a s u b cl u st e r a t a lower stage (25). A further
s u bcl us te r in E c onsi sts of items 34-15 'p iy e n o ' (go), ' erxome'

- 284 -
(come), 'fe v y o ' (leave), 'a ko lu Q o ' (follow ) and a f i n a l one includes
a p a i r o f items 29 and 11 'p e rp a ta o ' (walk) and 'tre x o ' ( run) which
are j o i n e d a t a lower stage (24) by ' p i f t a o ' (jump).

The m a t r i x has all the co-occurrences of the 34 items tested. As

a l re ady noted, both the content o f the HCS and the co-occurrences of
the m a t r i x need t o be considered c a r e f u l l y .

e. I n te r p r e ta tio n of results

On the basis o f the c l u s t e r s in the HCS and s u b j e c t s 7 comments on the

individual groups they formed, it can be s a f e l y argued t h a t in the
absence o f any c ontext ual (or o t h e r ) information, a ll the verbs of
the set are mainl y understood as referring specifically to human
motion. The 'most g e n e r a l ' meaning i deal i s not o p e r a t i v e , in t h i s
SST a t l e a s t .

The following items have both a c au sa ti ve and a non-causati ve

interpretation: 'k s a p lo n o ' (lie down), 'y o n a tiz o ' (kneel), 'k ila o '

(ro ll), 'y lis tra o ' (slide, slip ), 'v u lja z o ' (sink), 's triv o ' (turn),
'p e ta o ' (fly , throw), 'tre x o ' (run), (see 3.2) With the notabl e
exception of 'p e ta o ', none of the remaining items appears in the
causati ves cluster. This may imply that the non-causati ve
understanding i s the most prominent one, which c o n s t i t u t e s a d d i t i o n a l
evidence in f av our of deriving such causati ves from their non­
c au sa ti ve counterparts. The two interpretations of 'p e ta o ', i.e.
'fly ' and ' t h r o w ' , although they have in common something l i k e 'move
in the a ir', are evidently kept distinct. Almost half of the
subj ects (18) put 'p e ta o ' together with 'p ifta o ' (j ump) , i.e. as a
non- ca us at iv e, while it also has 22 and 20 co-occurrences wi th
'rix n o ' (throw) and 'tin a z o ' (shake up) respectively, i.e. as a
causative. Besides 'p e ta o ', 's t r i v o ' (turn) has a c a us a t i ve and a
non-causati ve interpretation, which are equally common. It is
perhaps noteworthy that it is almost never clustered t o g e t h e r wi th
the causati ves o f t h i s s e t .

- 285 -
It seems that human motion is mainly understood as 'in te n tio n a l'/
'agentive'. Wi thi n the E c lu s te r three c h a r a c t e r i s t i c a l l y agent ive
verbs: 'tre x o ' (run), 'p e rp a ta o ' (walk), 'p /5ao' (jump) join the
cluster at a l ower stage, but this can receive a number of
interpretations not directly related to agentivity, but rather to
p r o p e r t i e s such as ' l e g s as i ns tr ument ' or ' manner'.

Another t y p i c a l l y a gent ive verb, namely 'a k o lu d o ' (follow ), is p a r t

o f the s u b - c l u s t e r o f items 34- 15 which are not n e c e s s a r i l y a g e nt i ve .
The most unexpected clu stering in this r espect i nv ol ve s ' sernome'
(creep, crawl), which belongs to the group characterized by most
subj ects who put its constituent items together as i n v o lv i n g
'ac cid e n ta l', 'negative', ' u n i n t e n t i o n a l ' motion. As a l r e a d y pointed
out , 'sernom e' cannot be 'unintentional', in the sense that it
r e qu i re s (intentional) motion of arms, l e g s, etc. The obvious
exp l an a ti o n is that it is quite atypical of human CL and is
associated with 'undesirable' situations. This, combined wi th
something l i k e ' ma jor p a r t of the body in contact wi th ground' can
e x pl a i n the unpredi cted result, i.e. its appearing together wi th
items which are very low on an a g e n t i v i t y s ca l e, but share the above
properties with it. No g e n e r a l , traditional classificatio n that I
know o f ( i n c l u d i n g the one attempted in 4 . 2 ) accounts f o r t h i s piece
o f i n fo r m a t i o n .

N ot i c e , further, in this r e sp e ct , that 'in te n tio n a lity ' or

'a gentivity' is not in the least ignored, and t h a t it i s not an
independent property. The only verb in the whole set necessarily
involving 'w ater' as 'medium' namely 'v u lja z o ' (sink), is also p a r t
of the 'negative - accidental' cluster C. Unintentional or 'non-
agentive' motion i s l i n k e d wi th properties like 'accidentality' and
'undesirability' and i s o f t en combined wi th 'major p a r t o f the body
in c onta ct with e n vi r onm ent '. It can be also shown that it is
somehow combined wi th 'downward' motion.

C l u s t e r B contai ns ' v e r t i c a l ' motion non-c aus ati ve s. The s al i e n c e of

'vertical' direction has been pointed out more than once. Notice
that there exist co-occurrences of items sharing a pr op e rt y of

- 286 -
'upward' or 'downward' direction and d i f f e r i n g even in c a u s a t i v i t y ,
e.g . ' ipso no ' (raise) - ' aneveno' (ascend) -> 13, ' ipsono' (raise) -
' ska rfa lo n o ' ( c l i mb ) -> 12, 'k a te v e n o ' (descend) - 'g re m iz o ' (pull
down) -> 8. These scores are much lower than the ones corresponding
to the p a i r s which share the pr op e rt y of ( non- ) causativity. This
w ill not be considered here as evidence in favour o f the s a l i e n c e o f
'causativity' vs 'v e rtic a lity ' or some s p e c i f i c vertical direction,
c on t r a r y to M i l l e r ' s (1972) p r a c t i c e c r i t i c i z e d in 5 . 1 . 1 . What these
results show i s t h a t the combination o f v e r t i c a l CL w i t h absence or
presence of causativity is ext remel y salient. Hence: ' aneveno'
(ascend) - 's k a rfa lo n o ' ( c l imb ) have 39 co-occurrences although they
d iffer in the specification of 'manner' and 'instrument' and
' aneveno' (ascend) - 'kateveno' (descend) score 35, although the
former is 'upward' and the latter 'downward'. Sim iliarly, in the
causati ves c l u s t e r D, 'anevazo' ( t ak e up) - ' ip s o n o ' (raise) have 48
co-occurrences, 'a n e va zo ' - 'kre m a o ' (hang) score 31 and ' ipsono' -
'krem ao' -> 30.

The last 2 pairs mentioned in connection wi th v e rtic a lity and

c a u s a t i v i t y pr ovi de a f u r t h e r pi ece o f unpredicted i n f o r m a t i o n . The
verb 'k re m a o ' (ha ng) , far from being understood in i t s most general
sense which i nvolves only 'change of point-of-support' (see List
V II), is obvi ousl y fe lt to i n vol ve 'upward' CL. This can be only
e xpl ai ned i f the p r o t o t y p i c a l i nstances of i t s a p p l i c a t i o n are taken
into consideration, linked with hanging c l o th e s on a h a t - s ta n d and
p i c t u r e s on the w a l l . All these i nv ol ve t y p i c a l l y r a i s i n g the o b j e c t
i n quest ion. This leads i n d i r e c t l y to the conclusion t h a t c l u s t e r A
items are probably not understood as i n v o l v i n g 'change o f p o i n t - o f -
support' either, but r a t h e r 'change o f p o s t u r e ' . Thi s i s r e i n f o r c e d
by s u b j ec t s ' comments on the content of A (or some p a r t of it),
characterizing it as 'different p os i ti on s of the human body'. As
al re ady poi nted out in 4 . 2 t h i s i s c e r t a i n l y not the 'most g e n e r a l '
understanding o f t he verbs in que st i on, namely ' k s a p l o n o ' ( l i e down),
'kaQome' ( s i t down) and ' y o n a t i z o ' ( k n e e l ) .

The d i s t i n c t i o n between CP and CL i s not i gnored, although in the

absence o f any CP verbs which are not also 'change o f p o s t u r e ' , it is
i mpossible to decide which p a r t i c u l a r pr op e rt y i s taken i n t o account

- 287 -
when they are clustered together. What is worse, all the CP non­
caus ati ve s o f the set o f f e r e d t o subjects i n v ol ve 'downward' motion
as w e l l , w it h the exception of 'sikonom e' (rise). Now this last
item can be e i t h e r CL or CP. In e i t h e r case it i nv ol ve s 'upward'
motion. It i s perhaps i n d i c a t i v e to n o t i c e t h a t ' s i k o n o m e ' is judged
as much c l o s e r to the CP group, as t h i s is the interpretation it
r ec ei ve s when a p p l ie d to human motion. Not ic e t h a t it has 30 co­
occurrences w it h 'ka&ome' (sit down). This i s the hi ghe st score f o r
'sikonom e' and may imply t h a t it is understood as i t s antonym. It
also has 28 and 26 co-occurrences wi th 'y o n a tiz o ' (kneel) and
'k s a p lo n o ' (lie down) respectively. The combination of 'v e rtic a l'
motion, CP and absence o f c a u s a t i v i t y can e x pl a i n these scores. It
i s important to n o t i c e , however, t h a t these p r o p e r t i e s are not in the
l e a s t independent i n t h i s cluster. 'Change o f pos tur e' is t y p i c a l l y
'vertical' and 'non-causative' (si nce all these verbs are mainly
understood as involving 'human body position'). Combinations of
p r o p e r t i e s are o ft e n b e t t e r e x p l a i n a b l e in terms o f t y p i c a l 'scenes'
or ' f ra m es ' than in terms o f most general understandings o f i s o l a t e d
items. In practical terms, 'sikonom e' (rise) can be easily
i n t e r p r e t e d as p a r t o f a scene where somebody who is seated, kneeling
or lying down, 'ris es ' (stands up). 'V e rtica lity' and 'absence of
causativity', although ext remel y salient, as a lr e a d y shown, cannot
e xp l a i n the behaviour of 's ik o n o m e '. It has only 17 and 13 co­
occurrences with 'aneveno' (ascend) and 's k a rfa lo n o ' (cl imb)
respectively, although they a l l share 'upward' motion (and not j u s t
' v e r t i c a l i t y ' ) , CL and absence o f c a u s a t i v i t y . These d i f f e r e n c e s are
only e x p l a i n a b l e if 'sikonom e' is t y p i c a l l y interpreted as implying
'change o f p o s t u r e ' .

There are two verbs in the s et offered to s u bj e c t s , 'S ia s x iz o '

(traverse) and 'p e rn a o ' (pa ss) , which were c a t eg or i z ed under ' p a t h '
( or ' p a ss ag e ') in 4.2. They appear t og e t h e r 39 t im es , for 10 of
which they form a separate group. In this sense, the relevant
pr ope rt y may be said to have some psychological validity, at least
when combined w i t h absence o f c a u s a t i v i t y . What is more i n t e r e s t i n g ,
they are both c l o s e r t o the 'random wal k' verbs, i.e. 'p e rife ro m e '

and ' t r i y i r i z o ' (roam around) w i t h which they have 25 co-occurrences,

than to the remaining non-causati ves in the same c l u s t e r E. This may

- 288 -
be interpreted as implying that 'absence of direction and goal',
which is a necessary p r op e r t y of the 'random w al k ' group and a
prototypical one of these verbs has some validity. Alternative
e xpl ana ti ons are also p o s s i b l e . Not ice t h a t ' p e r n a o ' (pass) i s oft en
used as 'pass by' or 'drop i n ' . A scene i n v o l v i n g roaming around and
passing by vari ous places while doing so may be r es po ns i bl e for
linking ' p e r n a o ' w i t h the 'random wal k' ver bs. ^

There i s one example o f 'change o f o r i e n t a t i o n ' or ' r o t a r y m ot i o n ' ,

namely 's triv o ' (turn), which i s the most i s o l a t e d verb o f t h i s s et ,
as i t appears 22 times by i t s e l f . The suggestion made i n 4 . 2 . 3 t h a t
change of orientation may be linked wi th 'random walk' is not
compl etel y unwarranted on the basis o f t e s t r e s u l t s , as 's triv o ' has
15 co-occurrences with the 'random wal k' verbs. Its 15 co­
occurrences wi th 'b ia s x iz o ' (traverse), however, are not e x p l a i n a b l e
on the basis o f the a n al y s i s in 4 . 2 . 3 .

There is further one example of 'dependent motion', the verb

'a ko lu Q o ' (follow) which is also fa irly iso lated (15 times by
its elf). This is not sufficient evidence in f avour o f postulating
'dependent motion' as a separate category (as suggested in 4 . 2 . 2 ) .
Although 'a ko lu Q o ' is the only verb in this set which combines
'motion a f t e r somebody/something' l i n k e d w i t h strong 'intention' (as
suggested in 4.4), it is fa irly wel l integrated within the sub­
cluster of E consisting of items 3 4 - 15 , namely: 'p iy e n o ' (go) ,
'fe v y o ' (leave), 'e rx o m e ' (come). This is not explainable on the
basis o f the a n a l y s i s o f f e r e d in 4 . 2 . An a l t e r n a t i v e i n t e r p r e t a t i o n
seems much more p l a u s i b l e , which i s based on the f a c t t h a t 'a ko lu Q o '

has 25 co-occurrences ( t h e maximum score f o r t h i s verb) w i t h 'erxom e'

(come). It is quite probable t h a t 'a k o lu Q o ' (follow) is typically
linked with 'someone going purposefully after someone e lse', i.e.
' goi ng towards some o t he r p e r so n ' . These c on di ti o ns seem t o i nvolve
different 'directional adverbials' in formal terms. Appar entl y the
' t o wa rds ' vs ' a f t e r ' distinction i s not o f primary importance, w hi le
'having as a goal the p o s i t i o n o f some ot he r person' is, and t h i s i s
p r e c i s e l y what ' a k o l u Q o ' and ' e r x o m e ' (come) have in common (compared
to the remaining verbs o f the s e t ) . This gives an idea o f what the
prototypical understanding of both verbs may involve, and bears

- 289 -
directly on the important issue of deixis, to be discussed
immediately below.

Wi thi n c l u s t e r E the s u b - c l u s t e r mentioned in the preceding paragraph

includes an antonymous d e i c t i c p a i r 'p iy e n o ' (go) and ' e r x o m e ' (come)
which any analysis would consider together. The number o f co­
occurrences of these verbs is sig n ific a n tly high: 39. N ot ic e ,
however, that 'fe v y o ' (leave) also has 38 co-occurrences wi th
'e rxom e' and 37 w i t h 'p iy e n o '. We a c t u a l l y get a trip let, rather
than an i s o l a t e d antonymous p a i r , as is the case w i t h 'b e n o -v y e n o '
(49 co-occurrences) for i ns t an c e. The trip let is explainable in
terms o f 'tra d itio n a l' properties. All t hr ee verbs are l ess marked
(involve fewer specifications) than the remaining CL non-causatives
o f the s e t . What i s not e x p l a i n a b l e in such terms i s the high number
of co-occurrences of 'e rxom e' wi th 'a k o lu Q o '. In terms of the
categories of ' p l a c e and person d e i x i s ' , 'p iy e n o ' and ' e r x o m e ' would

appear t o be e q u a l l y deictic. I t seems, however, c l o s e r t o the f a c t s

to accept that 'e rx o m e ' (come) i nvolves typically 'motion to the
location of a person' unlike 'p iy e n o ', to which the most general
understanding of deixis involving 'place or person' is more
appropriate. Individual listings show that when 'p iy e n o ' and
'e rx o m e ' are l i s t e d separately, t he latter verb is grouped with
'a ko lu Q o ' (follow) which also i n vol ve s 'movement towards a pe rson' .
On the o t he r hand, 'p iy e n o ' (go) is, in such cases, listed t oge the r
with 'fe v y o ' (leave) and 'p e rp a ta o ' (walk). The most immediate
interpretation of these data is that 'p iy e n o ' its e lf is sometimes
used to imply 'fe v y o ' (I am l e a v i n g ) . Finally t he co-occurrences
w it h 'p e rp ata o ' (wal k) point to the direction of t he prototypical
understanding of this last ver b, which although specified for
'manner' (instrument, medium, e tc .), descr ibes the most
c h a r a c t e r i s t i c way o f CL or ' d i s t a n c e c o ve r i n g' for humans ( c l o s e l y
fol lowed by ' t r e x o ' (run)). In t h i s sense it is closer to the
' ge ne ral motion CL non- c au s at iv e s' than other 'manner' s p ec i f y i ng
verbs. The distinction between 'general' and 'manner' s p ec i f y i ng
verbs is not in the least indisputable and should be considered
c a r e f u l l y in connection wi th the SST r e s u l t s .

- 290 -
Some o f the verbs discussed in 4 . 2 . 4 under 'manner' are si ngl ed out
by M i l l e r and Johnson-Laird (1976:550) as d e sc r ib i ng 'main global
locomotory m o t i o n s ' . Wi thi n the set given to s ubj ects such MGMVs
ar e: 'p e r p a t a o ' (walk), 'tre x o ' (run), 'sernom e' (creep, crawl),
' s k a rfa lo n o ' (climb). As can be e a s i l y attested in the HCS these
verbs do not form a cluster, despi t e the fact that they share a
number o f p r o p e r t i e s . They are a l l a ge nt i ve , non -c a us at iv e , 'manner'
specifying, CL and 'distance cove ri ng' in particular, imply 'body
i n vo l v e me n t ' , and 'contact with ground of extremities of lim bs'.
Most traditional analyses consider them t o g e t h e r , i n c l ud i n g M iller
and Johnson-Laird ( i b i d . ) and 4 . 2 . 4 o f the present study. A look a t
individual p a i r i n g s o f verbs in the m a t r i x r e v e a l s t h a t the c l ose st
pair is 'p e rp a ta o ' (wal k) - 'tre x o ' (run), wi th 35 co-occurrences.
This i s e x p l a i n a b l e both in t r a d i t i o n a l and p r o t o t y p i c a l terms. The
next best p a i r i n g involves 'p ib a o ' (jump) which appears 24 times wi th
'tre x o ' (r un) and is not i ncluded in the 'main global locomotary
motions' set. In 4 . 2 . 4 'p ib a o ' (jump) was considered as t y p i c a l l y
' n on - di st an c e c o ve r i n g' and as i n v o l v i n g some element o f 'ene rgy' or
'fo rc e ', a common pr o pe r ty w i t h 'tre x o ' (run). It also has 20 co­
occurrences with 'p e rp a ta o ' (walk), however, an i n d i c a t i o n t h a t the
m ajo rity of the properties mentioned as c h a r a c t e r i z i n g the 'main
global locomotory motions' set are also a p p l i c a b l e in t h i s case and
are not i gnored. What does not show immediately in the HCS but can
be checked in the m a t r i x is that 'p ib a o ' has 18 co-occurrences wi th
'p e ta o ' (fly ) and 15 w it h 's k a rfa lo n o ' (climb). The conclusion t h a t
'upward' motion is also part of its specification is fa irly
inevitable, as i s a lso the r e c o g n i t i o n o f the f a c t t h a t in the case
of 'p ib a o ' (jump) properties other than 'v e rtic a lity ' are fe lt as
more salient. Ne ve rt he l es s, the main problem w it h the set in
question does not lie in t he excl usion of 'p ib a o ' but in the
inclusion of ' sk a rfa lo n o ' ( c l imb ) and 'se rno m e ' ( c re ep , crawl). As
a lr e a d y noted both verbs are quite removed from the ones just
discussed, as 's k a rfa lo n o ' is c lu s t e r e d together w it h 'v e rtica l'
motion non-causati ves (cluster B) and whi le 'se rno m e ' is p a rt of
cluster C, which i ncludes non-causatives marked for absence of
intentionality ( ' u n d e s i r a b l e ' m ot ion) , ' c o n t a c t o f major p a r t o f body
w i th e nvi r onme nt ', and perhaps also some 'downward' element, which i s
suggested in 4 . 4 to be c l o s e l y l i n k e d w it h absence o f a g e n t i v i t y .

- 291 -
The question therefore arises to what extent t he 'g eneral' vs
'manner' distinction advocated in 4 . 2 is at all plausible. The SST
results show t h a t although i t i s not ignored, it is not very s a l i e n t
either, compared to ot he r properties, nor is it independent of
p r o p e r t i e s such as 'medium', 'instrument' and t y p i c a l CL f o r humans.
If the distinction were not recognized by s u bj e c t s , 'p e r p a t a o '
(walk), 'tre x o ' (run), 'p ib a o ' (jump) would not have c o n s t i t u t e d a
clear sub-cluster of E (CL non- causati ves, 'indeterm inate'
direction). On the o t he r hand, the co-occurrences o f 'p iy e n o ' (go)
(a most 'general' motion verb) w it h those verbs of the set which
specify 'manner' show, once agai n, the inadequacy of general
classifications and the importance o f p r o t o t y p i c a l images. Compare
these co-occurrences:

perpatao (walk) -> 21

trexo (r un) -> 11
pibao (jump) -> 4
piyeno ska rfa lo n o ( c l imb ) -> 4
(9 0 ) sernome ( c r e e p , c r a w l ) -> 4
petao (fly , throw) -> 2
k ila o (ro ll) -> 2

These s i g n i f i c a n t d i f f e r e n c e s can be expl ai ned only i f one considers

t h a t the p r o t o t y p i c a l image o f 'p iy e n o ' (go) i nvolves something l i k e
' i n t e n t i o n a l d i s t an c e covering on f o o t ' . The l a s t and the f i r s t p a r t
of this rough specification ('in ten tio n a l' and 'on fo o t') are
i nt er dep ende nt . This s p e c i f i c a t i o n i s also a p p l i c a b l e t o 'p e rp a ta o '

(walk), i n v o l v i n g t he most t y p i c a l way o f CL f o r humans, as well as

to 'tre x o ' ( run) which is, however, less characteristic/typ ical in
t h i s r e sp e ct . None o f the remaining 'manner' s p e c i f y i n g verbs o f the
set involves typical CL or di s ta n ce covering for humans. This can
explain t he remarkably low scores they have w it h 'p iy e n o ' ( g o) .
Additional data as t o how these verbs are c l u s t e r e d and f o r what
reasons have been a lr e a d y discussed and need not be re pea te d.

Wi thi n c l u s t e r D, which i ncludes CL c aus a ti ve s , two s u b - c l u s t e r s are

identified in the HCS. The one which consi sts o f items 1 7- 14, i.e.

- 292 -
' ip so n o ' (raise), 7a n e v a z o ' (take up), 'kre m a o ' (hang), involves a
combination of 'upward7 motion and presence of 'accompaniment'.
Items 16-23, i.e. 'rix n o ' (throw), 'tin a z o ' (shake up), and ' g r e m i z o '
( p u l l down) are more c l o s e l y l i n k e d t o the c au s at iv e understanding of
'p e ta o ' (throw), w i t h which 'rix n o ' (throw) and 'tin a z o ' (shake up)
have 22 and 21 co-occurrences respectively. These i n vo lv e 'absence
o f accompaniment' and ' i m p e t u s ' , which are not independent p r o p e r t i e s
since in the case of CL causati ves ' i mpetus' i m pl ie s 'absence of
accompaniment'. It does not therefore make sense to talk about
r e l a t i v e s a l i e n c e o f a s i n g l e pr ope rty in t h i s case e i t h e r .

The tendency of subj ects to form p a i r s is quite noticeable in the

HCS. Some o f these p a i r s were a lr ea dy discussed. Of the remaining
ones, 'beno ' (e nte r) and ' v y e n o ' (exit) have 49 co-occurrences (i.e.
only one s ubj ec t separated them), w hi le 18 times t hey are l i s t e d in
isolation. This may imply t h a t some pr ope rty involving 'two/three
dimensional a r ea ' is f a i r l y salient, provided no o t h e r p r o p e r t y ( i e s )
o f those a l re ady mentioned as very salient is also i nv o lv e d. This
must be noted, in or der to account for the fact that 'v u lja z o '

(sink), which also involves 'motion into a three -di mens iona l area',
has only 2 co-occurrences w i t h 'b e n o '.

The two p a i r s o f near-synonyms o f t h i s s e t , i.e. 'a n e va zo ' ( t ak e up)

- 'ip s o n o ' (raise) and 'p e rife ro m e ' - 'triy iriz o ' (roam around) were
included in or de r to check whether the '-om e' mediopassive ending
would pl ay a r o l e in s u b j e c t s ' c l a s s i f i c a t i o n s and whether 'a n eva zo'

would have more l i s t i n g s w it h i t s non-causati ve c o u n t e r p a r t 'aneveno'

(ascend) than 'ip s o n o ' (raise) would. The results show that
morphological cues ( ' ~ome' ending) played no role in s u bj e ct s '
judgments and that s ubt le differences of meaning {'ip s o n o ' -
7a n e v a z o ' ) do not a f f e c t the o v e r a l l c l a s s i f i c a t i o n .

The f a c t t h a t 'fe rn o ' ( b r i n g ) has only 2 co-occurrences w it h 'erxom e'

(come) could have been interpreted as implying that their common

deictic component i s in fin ite ly less s a l i e n t than the c aus at ive one
(which the two verbs do not s ha r e ) . It seems, however, more
plausible to assume that if the pr oper ty of causativity ( or its

- 293 -
absence) is a shared one, subj ects tend t o form p a i r s o f antonyms or
near-antonyms and do not need to bother about which particular
properties are shared unless a very salient one i s i n vol ve d. This
expl ai ns why 1f e r n o ' (bring) and ' s t e l n o ' (send) are such a ' s t r o n g '
pair (48 co- occurrences, 30 o f which are unaccompanied by any ot he r
item), although the exact antonym of 'fe rn o ' (bring) is actually
' p/ y en o^ ' (take to ).^ It also e xpl ai ns why 'beno - vyeno' is an
e q u a l l y strong p a i r , while 'aneveno - kateveno' is l ess strong: the
l a s t case is the onl y one i n v o l v i n g the extremel y s a l i e n t p r o p e r t i e s
of 'upward' and 'downward' motion; besides, the p o s s i b i l i t y e xi s t e d
of l i s t in g e i t h e r member o f the antonymous p a i r w i t h o t h e r verbs o f
the set shari ng this property w h i le not being members of o the r
antonymous ( or synonymous) p a i r s .

To sum up, it seems unwarranted t o draw any conclusions bearing on

the overall organization of a semantic domain on t he basis o f the
data from any subset. The same a p pl ie s to de ci si ons on the exact
nature o f p r o p e r t i e s or dimensions which seem to oper at e w i t h i n the
subset. S p e c i f i c s t r a t e g i e s o f subj ects performing SSTs, such as the
tendency to form p a i r s o f (ne ar ) synonyms and ( near ) antonyms, pl ay a
decisive r o le , as does also the actual choice of v a r i a b l e s (items).
The p o s s i b i l i t i e s o f combinations open to subj ects have t o be taken
seriously into account. A further important f a c t o r emerges, namely
that a number of properties are not independent. In view o f all
these, interpreting SST results in terms of number of shared
attributes/properties (e.g. Long 1975) or r e l a t i v e salience between
isolated properties (e.g. M i l l e r 1972) does not seem wel l founded.

Less strong conclusions can, however, be drawn. The most i mportant

among these seem to be the r o l e of the r e l a t i v e salience of certain
combinations o f p r o p e r t i e s in semantic s i m i l a r i t y judgments and the
psychological valid ity of specific, prototypical understandings of
c at e g o r i e s as opposed to most general ones.

- 294 -
5.2 P r o t o t y p i c a l i t y t e s t s w it h Modern Greek motion verbs

Taxonomic relations between terms and the problems raised in

connection w it h verbs, and MGMVs in particular, are discussed at
length in 4 . 1 , which contains also an a nal ys is o f such r e l a t i o n s in
the f i e l d under c o n s i d e r a t i o n . Section 4 . 1 . 2 concentrates on how to
determine which items are ' b a s i c l e v e l ' in t h i s field and e la bor a te s
on the notions of 'lin g u istic markedness' and 're la tiv e class
inclusion'. The present s ecti on focuses on the impor tant principle
el abor at ed by Rosch and her col leagues (e.g. Rosch and Mervis 1975,
Mervis and Rosch 1981) that subordinate c at e g o r i e s are not
equidistant from the i n c l u s i v e category name, i.e. that all members
(of a c at egor y) are not equally representative of their category.
Evidence in f av our of the non-equivalence of cat egor y members was
firs t offered by Berlin and Kay (1969) where an illuminating
distinction was drawn between f ocal and non-focal col our s. The
former c o n s t i t u t e poi nts in the c ol our space which speakers o f q u i t e
different languages judge as most r e p r e s e n t a t i v e (i.e . as the best
examples) o f the 11 basic col our c a t e g o r i e s .

Rosch and her col leagues have shown t h a t gr a di en t s o f r e p r e s e n t a t i v e s

are also found in many semantic c at e gor i es and i n p a r t i c u l a r both in
those constituting biological taxonomies (ani mal s, plants) and in
those belonging to non-b iol o .gical ones, such as artefacts
(furniture, musical i n s t r u me nt s ) . Rosch’ and Mervis (1975)
demonstrate t h a t category members d i f f e r in the e x t e n t to which they
share attributes with ot he r category members. This v a r i a b l e they
call 'fam ily r esembl ance', following Wittgenstein (1953). Examples
which have most a t t r i b u t e s in common wi th o t he r items w i t h i n the same
category share few (if any) attributes wi th members of ot he r
(sim ilar) categories. These examples have the hi ghe st family
resemblance scores and are judged as the most representative or
prototypical o f t h e i r i n c l u s i v e c at egor y.

Pulman (1983) investigates whether verb c a t eg or i e s also show the

pr otot ype effect and whether, in case they do, the same p r i n c i p l e
(i.e . f a m i l y resemblance) can be held responsi bl e f o r the formati on

- 295 -
of pr ot ot ype s. His investigation is of immediate concern to the
present study, as a number o f the observations he makes are also
a p p l i c a b l e t o the f a c t s o f MGMVs. N ot ic e , f o r a s t a r t , t h a t although
Pulman was drawing m a t e r i a l from a l l d i f f e r e n t verb areas o f English,
he ended up w i t h only 8 hyponymic sets c onsi st ing o f a hypothesized
basic l e v e l and s i x hyponyms. His remark " i t turned out t o be q u i t e
d ifficu lt to f i n d enough basic l e v e l verbs w i t h a s u f f i c i e n t number
o f hyponyms" ( i b i d . : 1 1 0 ), i f a p pl i e d to the present study o f a s i n g le
verb domain (MGMVs), would be something o f an understatement in view
o f the d i f f i c u l t i e s poi nted out in 4 . 1 . An a d d i t i o n a l problem, which
has emerged at v ar ious poi nt s throughout this study, is the
particular sociol i n g u i s t i c compl icati ons faced by anyone conducting
lexical research in MG.

In view o f the aforementioned d i f f i c u l t i e s , the p i l o t t e s t I carried

out in order t o see whether the protot ype e f f e c t obtained f o r MGMVs
contained 11 groups o f verbs, some o f which were extremel y u n f a m i l i a r
and/ or belonged to different registers, constituted dubious cases,
were cross-classified or were marked SI or S3. The test asked
subj ects for subjective ratings of typicality/representativeness
along a 7-point scale, following Rosch and Pulman. Twelve f o ur t h
y ea r undergraduates of the English Department of Athens University
served as subj ects f o r the p i l o t test. They were given 11 sets of
MGMVs, each c o n s i s t i n g o f a basic l e v e l term ( i n my view) and 7-10
hypothesized hyponyms, and were asked to decide which member of each
category was more c h a r a c t e r i s t i c or t y p i c a l o f t h a t c at egor y.

Subjects were instructed to extend the ' s ta n d a r d ' (in Rosch and
Pulman's t e s t s ) 7-point scale t o 10 poi nts f o r those sets that had
more than 7 hyponyms, i f they f e l t i t was necessary. They were also
instructed to l eave out items about which they had doubts (i.e .
because they e i t h e r di d not know t h e i r meaning or did not know how to
c l a s s i f y them). The idea was t h a t i f c e r t a i n items were c o n s i s t e n t l y
left out they would not be i ncluded in the f i n a l test. Apart from
that the instructions were in essence the same as those given by
Rosch and Pulman but somewhat more d e t a i l e d . ^ The i n s t r u c t i o n s are
included below ( t r a n s l a t e d i n to Engl ish) and fol l owed by the w r i t t e n
example o f what subj ects were supposed to do.

- 296 -
 "You have 11 groups o f verbs which s i g n i f y 'movement'. Each
group consi sts o f one verb which i s considered the basic one o f
the group and 7 t o 10 others which are considered hyponyms o f the
former, i . e . are c l a s s i f i e d under i t (and consequently i nclude i t
as p a r t o f t h e i r meaning).

The f i r s t group i ncludes 8 members o f the cat egory ' p e r p a t a o '

( w a l k ) , the second one 6 members o f the cat egory ' v i Q i z o m e '
( s i n k ) , and so on. You have to decide f o r each member o f the
category 'p e rp a ta o ', for i n st an ce , how characteristic
( r e p r e s e n t a t i v e ) o f t h a t category i t i s . The most r e p r e s e n t a t i v e
verb o f each cat egory i s t o be put i n box number 1, the l e a s t
r e p r e s e n t a t i v e in number 7. Only the l i t e r a l meaning o f these
verbs is taken i n t o c o n s i d e r a t i o n , not t h e i r f i g u r a t i v e uses
(e .g. 'm e ta fe ro ' ( t r a n s f e r , t r a n s p o r t ) is t o be understood as
implying ' a n tikim e n a' (objects) and not 'ta e k lo y ik a mu
b i k e o m a t a ' (my v oti ng r i g h t s ) ) .

Notice the f o l l o w i n g p oi nt s:

1. Use, as f a r as p o s s i b l e , a l l the boxes, that is try to

c l a s s i f y the verbs using the whole sca le.

2. A box may, n e v e r t h e l e s s , include more than one verb i f you

consider that these verbs are equally representative
o f the cat egor y.

3. For those groups o f verbs t h a t have more than 7 members you

may extend the scale up to 10, i f you wish.

4, I f you are i n doubt about the ' marki ng' o f a member of a

category or about i t s meaning, leave i t o u t . "

Example: skotono (k ill)

1. bolofono 2. e k t e l o 3. s f a z o 4.
murder execute massacre, s l a ug ht e r

5. Q isiazo 6. a f t o k t o n o 7.
sacrifice commit s u i ci d e

Two examples were also provided and shown on t he board, one of

i nc l u s i o n (hyponymic) r e l a t i o n s h i p s i n v o lv i n g the MG terms f o r :

- 297 -
 pi ant

tree flower bush

oak pine

and one o f p r o t o t y p i c a l i t y measurements wi th the MG terms f o r :

furniture
1. c h a i r , table 2. _____ 3. chest 4. _____
5. 6. h a t - s ta n d

The groups o f verbs which fol l owed the instructions were random as
was also the or der of the members w i t h i n each c a t eg or y . Subjects
were told that the experiment was a linguistic and not a
psychological one and t h a t the time they might need t o make t h e i r
judgments was o f no consequence. In view o f the d i f f i c u l t y repor ted
by the subj ects who had c a r r i e d out the 'semantic s im ilarity' task
wi th regard t o the instruction that they should try and take i n to
account all p os si bl e uses of the verbs ( provi ded they stick to
' p hy s ic a l m o t i o n ' ) , t h a t p a r t i c u l a r i n s t r u c t i o n was o mi tt e d.

A f t e r they had completed t h e i r answers, subj ects were encouraged to

discuss any problems they had had w it h the t e s t . On the basis o f the
di scussion which f o l l o w e d , it became q u i t e obvious t h a t a number o f
hyponyms would have t o be omitted in the final test, as they were
practically unknown to that particular and r a t h e r select group of
s u b j ec t s . This meant also t h a t some whole sets o f verbs were l e f t
w it h too few hyponyms and had to be excluded. Besides it became
clear that even sophisticated subj ects were unable to s p el l out
specific attributes f o r most verbs, although they could c o n f i d e n t l y
agree or di sagr ee w i t h proposed d e f i n i t i o n s and p oi n t out d i f f e r e n c e s
in the use o f r e l a t e d hyponyms.

- 298 -
The most important poi nts t h a t r e s u l t e d from the discussion were: (a)
the notions ' r e p r e s e n t a t i v e ' and ' c h a r a c t e r i s t i c ' which were e n t i r e l y
new t o the subj ects in this connection (p rototypicality) di d not
present a problem to anybody; and (b) the term ' d i s t a n c e ' was used
throughout the di scussi on, confirming Pulman's (ib id .) and my
intuitions that in making prototypicality judgments, s ubj ects are
actually estimating the degree o f semantic sim ilarity between an
inclusive category and each hyponym. This issue w i l l be taken up
again in a di scussion o f the f i n a l test results. At t h i s p oi n t a l l
t h a t needs t o be st at ed i s t h a t the actual r e s u l t s o f the p i l o t t e s t
were q u i t e encouraging concerning agreement on the best and the worst
examples a t l e a s t .

T he r ef o r e , the f i n a l t e s t was simply a m o d i f i c a t i o n o f the p i l o t one

along the f o l l o w i n g l i n e s :

(1) The gr oups/ sets of verbs were reduced to 8, since those sets
which presented unsurmountable problems were omi tt ed, except f o r
one which appears under C in the final test. That set was
included in order to check whether it would y i e l d equally bad
r e s u l t s wi th a l a r g e number o f s ubj ec ts .

(2) Items which proved to be 'unknown' to the p ilot test subjects

were o m i tt e d. Verbs marked SI and S3 were avoided (except for
one item 'a x Q o fo ro ' ( c a r r y baggage) in cat egory D) but verbs
belonging to d i f f e r e n t r e g i s t e r s were i ncluded. An example o f a
possi bl e object was i ncluded for the verb 'petj e/ne' (go for a
short t im e) o f set E, so t h a t the s p e c i f i c sense meant was made
clear.

(3) The in structions contained only poi nts 1 and 2 of those which
subjects' a t t e n t i o n was drawn f o r the p i l o t t e s t . The f i r s t one
was phrased more simply but the specification "the whole scale"
was u n d e r l in e d, as the p i l o t t e s t showed a remarkable r el uc ta nc e
o f subj ects t o use the whole scale i f less than 7 hyponyms were
i ncluded. For the same reason an a d d i t i o n a l item: 'e kso lo Q re vo '

(exterminate) appeared in box 4 o f the example pr ovi ded, whi le

- 299 -
 boxes 6 and 7 were f i l l e d w it h the items p r e v i o u s l y appearing in
boxes 5 and 6 r e s p e c t i v e l y .

The f i n a l t e s t was c a r r i e d out by 61 subj ects in two separ ate groups.

They were a l l native speakers o f Greek. One group consi sted o f 28
second y ea r undergraduates o f t he French Department and the ot he r one
o f 33 f i r s t y e a r undergraduates of the English Department o f Athens
University.

A couple of responses were incomplete but were not di scounted.

Despite my e f f o r t s , the f u l l 7 - p o i n t scale was not used when only 5
hyponyms were pr ovi ded. Following Pulman ( i b i d . ) a c h i - sq u a r e t e s t
was used to measure goodness of fit between observed and expected
distribution o f responses. The actual result is provided for each
ver b. Not ice t h a t f o r 6 degrees of freedom ( d f 6) and .05 l ev e l of
significance, the sig nificant val ue of the x^ t e s t is 1 2. 5 9. The
v as t m a j o r i t y o f the verbs t e s t e d were n e i t h e r evenl y nor randomly
distributed. A large number of the results obtai ned showed high
i n t e r - s u b j e c t agreement on what were the best and the worst examples
of each c at egor y.

In the presentation of results which f ol l ow s i mmedi atel y, rankings

and mean ratings are gi ven, the latter appearing under each
subordi nat e i tem. The r e s u l t s o f the x^ t e s t appear in parentheses
so that an accur ate picture is provided of the extent to which
s ubj ects agreed on the ranking o f each p a r t i c u l a r i t e m. ^

A. perpatao (walk)

r anking : 1 2
vim a tizo (pace) pezoporo (walk a long d i s t an c e )
mean r a t i n g : 1.606 1.623
x2 : (122.23) (121.3)

- 300 -
 3 4
s u la ts a ro ( s a u nt e r) b ra ske lizo (stride)
2.967 4 .1 64
(155.7) ( 9 6 .9 4 )

5 6
trik liz o (s t ag ge r ) parapatao (stumble)
5.721 6
(72.84) (74.64)

7
b u s u liz o (crawl (as of a baby))
6.197
(114.18)

Expectedly 'p e r p a t a o ' (walk) i s the category which y i e l d e d the best

results, (in the sense o f i n t e r - s u b j e c t agreement) p os si b l y because
owing to the p r o t o t y p i c a l i t y o f the a c t i v i t y , it i s easy to f i n d a
l a r g e number o f hyponyms well-known t o a l l subj ects and s u f f i c i e n t l y
'removed' from one another.

The items holding p o s i t i o n s 1, 2 and 3 i n h e r i t most o f the a t t r i b u t e s

considered ty pical of the category name (in 4.2 .4 ), i nc ludi ng
'reg u larity ' and 'continuity' o f motion. The f a c t that 'v im a tiz o '

(pace) does not typically i n vol ve ' re ac h in g a destination' (e .g.

'pacing up and down' is its most typical use) is evidently not
considered important enough to remove it from the firs t position.
Complete absence of goal, however, applicable to 's u la ts a ro '

(saunter), which is an i nstance of 'random walk', is probably

r esponsi bl e for its being fa irly far removed from the 2nd best
( 'p e zop oro ' ) .

Differences in register do not seem to pl ay an i mportant role:

' pezoporo' (walk a long d i s t a n c e ) is high register and 's u la ts a ro '

( s au nt e r) low register. Linguistic unmarkedness is not the most

- 301 -
important factor either: both 'trik liz o ' ( s t a gg e r) and 'p a ra p a ta o '

(stumble) are l i n g u i s t i c a l l y unmarked and q u i t e f a m i l i a r items (see

4.1), yet they are judged as poor examples of the cat egory name.
What is more i mpor ta nt , they are both more fr eq ue n t than 'p e zop oro'

(walk a long d i s t a n c e ) a t l e a s t .

It is t h e r e f o r e f a i r l y c l e a r t h a t the r e l a t i v e s al i e n c e o f s p e c i f i c
properties plays the most d e c i s i v e role in t h i s c at e go r y. The 4th
position is occupied by 'h ra s k e liz o ' (stride) which i s t y p i c a l l y an
'event' verb (no c o n t i n u i t y ) , the 5th and 6th p o s i t i o n s have verbs
d e s cr i bi n g 'irreg u lar' and 'd efective' kinds o f w a lk i ng , whi le the
least typical example is hardly an i nstance of 'p e rp ata o ' (walk).
The s impl est way of showing these differences in degree of
prototypicality in this case is through 's u b s titu ta b ility'
p ossibilities, in a common environment:

(1) 5 en p e r p a t a i k a n o n i k a *v im a tiz i, pezopori

s/he does not walk p r op e r l y s/he paces walks a long
d i s t an c e

? ? s u la ts a ri
saunters

trik liz i, parapatai

staggers

(2) hen p e r p a t a i a k o m i b u s u liz i

s/he does not (cannot) walk y e t s/he is c raw li ng (as o f a baby)

B. viBizom e (sink)

I 2
v u lja z o (sink) katapondizome ( s i n k to the bottom,
1.229 3.361 become inundated)
(225.4) (28.9)

- 302 -
 3 4
katabiow e (dive) vutao (dive)
3.934 4.623
(4.30) ( 1 9 .2 4 )

5 6
navayo (f ounder, fudaro ( s i n k t o the bottom)
4 .7 05 become s hi p- 4.951
( 2 4 . 7 6 ) wrecked) (1 7 .8 6 )

The only reason for having 'viQ Izom e rather than 'v u lja z o ' as the
i n c l u s i v e category here i s t h a t the former can be m a r g i n a l l y used f o r
intentional as wel l as u n i n t e n t i o n a l submersion into water. Quite
expectedl y ' v u lja z o ' (sink) i s almost unanimously judged as the verb
near est to the i n c l u s i v e category name.

One item 'k a ta b io m e ' (dive) fails to reach sig nificance and is
t h e r e f o r e di scar ded. There are also two i tems, those in p os i t i o n s 4
and 6 b a r e l y reach s i g n i f i c a n c e . T h e i r r a t i n g s are q u i t e unexpected.
The former one, 'v u ta o ' (dive) impl ies intentional and momentary
motion unlike the rest. It is a very fam iliar verb which is,
however, only partially included in 'viQ izo m e ' (sink). The l a t t e r
one 'fu d a ro ' (sink to the bottom) is low register and rather
unfam iliar, but it is ot herwi se a very good instance o f 'v iQ iz o m e '
(sin k).

It seems reasonable to assume t h a t the differences in 'manner' of

motion are not very distinct, neither are t he corresponding verbs
very f a m i l i a r . Subjects are t h e r e f o r e l e f t t o decide on the basis o f
presence or absence of 'in ten tio n ality' and 'fa m ilia rity ' which
emerges here as an important factor. Fam iliarity is not, however,
the only decisive factor here either. Not ic e t h a t 'ka ta p o n d iz o m '

( s i n k to the bottom, become inundated) i s not very f a m i l i a r , yet it

occupies the 2nd p o s i t i o n . Apparentl y i t s sense which does not imply
motion o f the o b j e c t (i.e . 'become i nu nd a te d ') is ignored here, at
least by some s ub j ec t s; once an item is given as a hyponym o f an
inclusive cat egor y, subj ects tend t o concentr ate on t h a t particular

- 303 -
use/sense it has which is d i r e c t l y associ ated wi th t he given hi gher
category name.

C. p ibao (jump)

1 2 3
s a lta ro (leap) anapibao (jump up tinazome (s pr ing up/
2 3.1 64 (and down)) 4.4 43 jerk)
(89.12) (13.03) (8.67)

4 5
petjem e (jump u p / j o l t ) ipe rp iba o (jump over)
4.590 4.705
(15.10) (22)

As a lr e ad y noted, this set was included d e s p i t e a l l evidence to the

c o n t r a r y on the basis o f the p i l o t t e s t r e s u l t s . Only 5 hyponyms are
included in this 'fin a l' v er si on of C and one of them 'tin a zom e'

( spr i ng u p / j e r k ) fails t o reach s i g n i f i c a n c e , w h i l e i t s near-synonym

'p e tje m e ' (jump u p /jolt) does not display any high inter-subject
agreement on how i t should be c l a s s i f i e d , neither does 'a n a p ib a o '

(jump up (and down)). As in the case of t he previ ous set, very

s l i g h t d i f f e r e n c e s are i n vol ve d, so i t makes no sense t o concentrate
on s a l i e n c e o f p r o p e r t i e s . Three cases are worth di s cu s si ng , namely
the items occupying p o s i t i o n s 1, 2 and 5.

It is interesting to n ot i c e that the items sharing almost a ll

important p r o p e r t i e s w i t h the category name, namely 's a lta ro ' ( l e ap )
and ' i p e r p i b a o ' (jump o v e r ) , occupy the f i r s t and the l a s t p os i t i on s
respectively. The former item i s normally understood as not implying
r e p e t i t i o n o f the jumping event ( i . e . a s e ri e s o f repeated jumps) and
i s used i n cases l i k e l e api ng over an o b s t a c l e , f o r i ns t an c e. It is
also low r e g i s t e r , lin g u istic a lly marked but quite fam iliar to the
speakers o f the age group t o which the subjects belonged. The l a t t e r
item i m p li es clearly jumping over something, is high register and
particularly infrequent and unfam iliar. The importance of
f a m i l i a r i t y and frequency in t h i s case is obvious.

- 304 -
The item in position 2, 'a n a p ib a o ' (jump up (and down)) has two
fa irly distinct uses: one implying 'jump as a result of being
startled' and another one implying 'jumping up and down'. The
subjects who c a r r i e d out the pilot test were mostly aware of the
l a t t e r use. This e xp l a i n s why ' a n a p i b a o ' f i g u r e s in the 2nd p o s i t i o n
r a t h e r than t o g e t h e r w i t h 'p e tje m e ' (jump up) and ' t i n a z o m e ' (spr ing
up): it is mainl y understood as implying a s e r i e s o f repeated jumps
and not as an i nstance o f CP (as i t appears in L i s t V I I ) .

D. p iye no ka ti kapu ( t ak e something somewhere)^

1 2
m etafero (transport, transfer) kuvalao (carry)
1.197 2.300
(2 4 5 . 6 8 ) (67.8)

3 4
metakomizo (move f u r n i t u r e ) proskom izo ( t a k e documents
4.322 4.590 t o an o f f i c i a l )
(9.94) (10.05)

5 6
axQoforo ( c a r r y baggage) m etavivazo (transmit)
4.656 5.328
(17.2) (29.8)

Only the two best and the worst examples o f t h i s category need to be
discussed as the remaining ones do not show i n t e r - s u b j e c t agreement.

The f i r s t t hi n g t o n o t i c e i s t h a t the d i f f e r e n c e s between subordinate

categories do not i n v ol v e ty pe s/ 'ma nne r' of motion but 'type of
object' caused t o move, f o r which the s uperordi nate cat egor y i s not
specified. Otherwise, 'accompaniment' is a shared a t t r i b u t e of a ll
the terms of the set. 'Indirect caus at ion' and th erefore the
possibility of appearing within predications which are low in
agentivity is onl y applicable to the inclusive category ( ' p iy e n o ')

- 305 -
and m a r g i n a l l y to the term judged as most prototypical of it, but
none o f the o t h e r s . Notice a l s o , that ' a x Q o f o r o ' ( c a r r y baggage) is
l i n g u i s t i c a l l y marked, u n f a m i l i a r and i n f r e q u e n t , while 'm e t a v iv a z o '
(transmit) is scarcely r e pl a c e a b l e by the category name, as the
o bj e c t 'caused to move' is typically abstract, e.g. 'a message',
'greetin gs'. In short, 'm e ta v iv a z o ' (transmit) i nvolves least
ty pical obj ect s 'caused to move' and although fam ilia r, it is
certainly of restricted use compared to the verbs appearing in
p o s i t i o n s 1, 2 (and 3) and is also o f high r e g i s t e r .

The best example 'm e ta fe ro ' (transfer) is e q u a l l y f a m i l i a r wi th the

second one 'k u v a la o ' (carry). They both imply t h a t the o bj e c t is
fa irly 'substantial' (it cannot be a l e t t e r , f o r instance), but the
l a t t e r item i m pl ie s also t h a t t he act i s unde si ra bl e ( e . g . t he o bj e ct
is heavy or the act is executed unnecessarily). In this sense
'ku valao ' (carry) is sligh tly more lin g u istically marked than the
item judged as most prototypical of this category.7 A further
d i f f e r e n c e between these two best examples i s the f a c t t h a t only the
former one is t y p i c a l l y understood as implying t h a t the o b j e c t must
reach a d e s t i n a t i o n , as does a lso the i n c l u s i v e category name. The
verb ' k u v a l a o ' need not imply any such t h i n g . E v i d e n t l y t h e r e is no
way o f choosing between the two a l t e r n a t i v e possibilities, i.e. of
tellin g whether linguistic markedness or salience of properties
determines the r e l a t i v e di stance of ' m e t a f e r o ' and ' k u v a l a o ' from the
i n c l u s i v e category name. Sim ilarly, i n view o f what has a lr e a d y been
noted wi th r espe ct t o 'm e t a v i v a z o ' (transm it), more than one f a c t o r
i s i nvolved in render ing i t the l e a s t t y p i c a l verb of t h i s cat egor y.

E. p iyeno (go)

1 2
porevome (go a long di stance proxoro (advance)
1.344 on f o o t ) 2.246
(213.7) (8 4 . 9 9 )

- 306 -
 3 4
petjem e (p.x. sto p e rip te ro ) ta ksibe vo (travel)
3.377 (go f o r a shor t t i me , e . g . 4.2
(77.86) to the kiosque) (17.17)

5 6
proelavno (advance ( m i l i t . ) ) parelavno (march)
5.852 6 .2 29
(60.62) (97.88)

This is another s et exhibiting remarkable inter-subject agreement,

comparable t o the 'p e rp a ta o ' (wal k) verbs and probably f o r the same
reasons. The i n c l u s i v e category c o n s t i t u t e s the most c h a r a c t e r i s t i c
'general' way o f coveri ng d i st a n c e and i t is easy t o find hyponyms
for it which are well-known to most subj ects and s u fficien tly
d i s t i n c t / r e m o v e d from one another.

No t ic e , firs t, that t he verbs occupying positions 2 and 4 do not

specify 'instrument' (on f o o t ) , l i k e the category name and u n l i k e the
remaining items o f the set. In view, however, o f what has al re ady
been noted in the previous s ecti on (SST results), the inclusive
category 'p iy e n o ' i s probably mainl y understood as involving 'going
somewhere on fo o t'. It is possi bl e that the same a p pl ie s to
'p ro x o ro ' (advance) and also t o 'p e tje m e ' (go f o r a shor t t i m e) owing
t o the s p e c i f i c environment ( t o the kiosque) suppl ied in t he t e s t f o r
the l a t t e r . T h e r ef o r e the onl y verb in the set which does not a ll ow
for such an understanding (i ns tr ume nt « feet) is 'ta k s ib e v o '

(travel). This may be the reason why t h e r e i s no g r e a t i n t e r - s u b j e c t

agreement as t o how t h i s verb i s to be c l a s s i f i e d .

Not ice f u r t h e r t h a t although both verbs judged as most p r o t o t y p i c a l

of this cat egor y are fam iliar, the firs t one is certainly l ess
fr e qu e nt than the second one. Besides, although 'porevome' (go a
long d i st an ce on f o o t ) may imply t h a t some d e s t i n a t i o n is going t o be
reached, such a s p e c i f i c a t i o n i s not a necessary one as i t i s f o r the
inclu sive c at e go r y. It is even l ess necessary in the case of

- 307 -
'p roxoro' (advance) which occupies the 2nd position. This means
either that this p r o pe r ty is not at all relevant, or that it is
understood as a shared pr o pe r ty o f a l l t h r e e verbs in quest ion. In
any case, the f a c t t h a t i t i s necessary in the case o f ' p e t j e m e ' (go
for a short t i m e) is not sufficient to render it closer to the
i n c l u s i v e cat egory than the two hyponyms a lr e a d y discussed.

The verbs judged as least typical of 'p iy e n o ' (go) are expectedl y
7p r o e l a v n o ' (advance (mil i t . ) ) which is evidently o f re s tric te d use
and much l ess fam iliar than all the rest and 'p a r e l a v n o ' (march)
commonly associ ated with ' wa lk ing in a parade', where the process
(and not the d e s t i n a t i o n ) i s i m por ta nt . This l a s t example i s also of
restricted use (therefore infrequent) but certainly more fam iliar
than ' p r o e l a v n o ' which i s judged as a b e t t e r i nstance o f ' p i y e n o ' .

In s h or t , a l l three fa c to rs , i.e. frequency, f a m i l i a r i t y and s al i en ce

of properties are again shown to pl ay a role in prototypicality
judgments.

F- fevyo ( l e a v e )

1 2 3
apoxoro (withdraw) anaxoro ( d e pa r t) aposirom e (retire)
1.817 2 3.1 97
(90.66) (82.23) (54.18)

4 5
apom krinom e (move away from) ksekubizom e (clear o ff)
3 . 23 4.47
(34.15) (42.82)

6 . 7
metanastevo ( emi gr at e) 5rapetevo (escape)
5.787 6.607
(78.55) (172.11)

- 308 -
This is the last set exhibiting 's pectacular' inter-subject
agreement. The reasons f o r t h i s need not be r e pea te d. The r e s u l t s
are easy t o di scuss, as a lot of i nf or ma ti on is suppl ied for the
verbs o f t h i s set i n 4 . 1 . 2 .

All three items occupying the firs t three positions are high
register. Although the di st an c e between the two best examples is
slight, it is worth noticing that it is the verb judged as l ess
representa tive than 7a p o x o r o 7 (withdraw) which actually covers the
same conceptual area as the inclusive category name. This is
probably due t o the fact that 7a n a x o r o 7 ( d e p a r t ) is l ess f re que nt
than 7a p o x o r o 7 ( wi t h dr a w ) , since Y e v y o 7 (leave) is commonly used
instead.

Not ice t h a t both best examples imply t h a t the act o f going away is
'fin a l', u n l i k e the verbs occupying the 3rd and 4th p o s i t i o n s . This
'non-complete di sappearance' is particularly characteristic of
'apom akrinome' (move away from) and i s probably re sp on s ib l e f o r its
being judged so f a r removed from the i n c l u s i v e cat egor y. Otherwise,
7a p o m a k r i n o m e 7 i s probably the commonest (most f r e q u e n t ) item of t h i s
set o f hyponyms and i s c e r t a i n l y l i n g u i s t i c a l l y unmarked.

Not ice that 7k s e k u b i z o m e 7 (clear off) covers roughl y the same

semantic area w it h Y e v y o 7 (leave) and is quite fam iliar to the
subjects. It is, however, evidently lin g u istically marked (strong
ne g at i ve connot at ion) and infrequent. It is a p p ar e n t l y for these
reasons t h a t i t occupies the 5th p o s i t i o n .

Both 7m e t a n a s t e v o 7 (e mi gr at e ) and 75 r a p e t e v o ' (escape) are f a m i l i a r

and lin g u istic a lly unmarked. The fact that they do not descri be
'simple7 physical motion but i nv ol ve also (promi nent) soci al
c h a r a c t e r i s t i c s is the obvious expl ana ti on why they ar e judged as the
least typical instances of 'fe v y o ' (leave). The risk of
underspecification i n case they are replaced by the cat egor y name i s
much greater than in the case of all the remaining verbs.
S ubstitutability and semantic di st an ce are obvi ousl y i mportant in
making prototypicality judgments. The salience of the specific

- 309 -
soci al connotati ons that go along wi th these two worst examples
cannot be doubted.

6. aneveno (ascend)

1 2 3
a n ifo riz o (go u p h i l l ) s k a r f a l o n o ( cl i mb ) ipsonome ( r i s e )
1.967 2.328 2.6 88
(73.26) (46.14) (38.28)

5
sikonome (rise, stand up) apoyionome (take o f f )
4.377 4.672
(21.05) (49.35)

ektoksevome (be launched) eksfenbonizom e (be hur led)

6.393 6.573
(123.15) (163.3)

The problems o f c l as s i n c l u s i o n a r i s i n g in t h i s set are discussed in

4.1.2. Not ice t h a t ' a n i f o r i z o ' (go u p h i l l ) and ' s k a r f a l o n o ' ( cl imb)
are most e a s i l y r e p l a c e a b l e by the i n c l u s i v e cat egory name, as they
both imply ' c o n t a c t w i t h supporting s u r fa ce ' and can r e f e r t o human
motion. It i s not t h e r e f o r e s u r p r i s i n g t h a t they occupy the 1st and
2nd p os i t i on s respectively. The s l i g h t di stance between the l a t t e r
verb and ' i p s o n o m e ' (rise) in the 3rd p o s i t i o n is r a t h e r unexpected;
as a l re ady noted 'ip son om e ' is most commonly associ ated wi th the
motion o f 'smoke' or ' d u s t ' ( r a t h e r than humans) and does not us u a l l y
imply 'contact with supporting surface' except in a f a i r l y 'fixed'
use, d e sc r i b i n g 'raising the fla g '. In all these cases ' ipsonom e'
(rise) is, however, r e p la c e a b l e by ' a n e v e n o ' ( the i n c l u s i v e category
name), which i s h a r d l y the case wi th any o f the remaining items.

The great di st an c e between the items occupying t he 3rd and 4th

positions, which r e c e i v e sim ilar translations, is h a r d l y s u r p r i s i n g .

- 310 -
As a lr e a d y noted in connection wi th the SST results, 'sikonom e'
(rise, stand up) i s t y p i c a l l y associated w it h a human body assuming a
standing position, and is, in this sense, quite dissim ilar from
/ ipsonom e' (rise).

It i s perhaps noteworthy t h a t ' apoyionom e' ( t ak e o f f ) can be e a s i l y

(and i s a c t u a l l y commonly) replaced by 'sikonom e' (rise), both used
to refer to airplanes in such cases. It is pos si b le that this
a ss o ci a ti o n br ings them fa irly close to one another . It is of
interest, moreover, that some subj ects put both items in the same
box, a f a i r l y uncommon p r a c t i c e in t h i s p a r t i c u l a r t e s t .

The two worst examples are a g r e a t d i st an ce from the r e s t and show

almost unanimous inter-subject agreement. A number o f factors are
i nvolved in t h e i r case, such as presence o f ' i m p e t u s ' , the f a c t t h a t
they are r a r e l y used f o r human motion and e s p e c i a l l y t h a t they are
quite infrequent. Needless to say, they are hardly replaceable by
the hypothesized i n c l u s i v e term.

H. pefto kato (fall down)

1 2
sorjazom e ( c o l l a p s e , come gremizome ( f a l l (in ruins))
1.459 crashing down) 2.885
(2 0 3 . 6 1 ) (60.62)

3 4 5
tubaro (overturn) k a tra k ila o (roll down) ku tru va la o (roll
4.574 4.820 5 . 01 6 down)
(51.42) (2 4.5) (17.17)

The category name here was supplemented w it h the s p e c i f i c a t i o n 'k a to '

(down), to r u l e out readings o f 'p e fto ' (fall) which are u nr e l a t e d to

the items s el e c t e d as i t s hyponyms f o r the present purposes. This
was perhaps an unnecessary s tep, as 'p e fto ' is probably typically
understood as (accidentally) fallin g down, in the absence of any

- 311 -
specification to the contrary (e.g. fallin g in tentionally on the
enemy).

All the items o f t h i s set are p r o t o t y p i c a l l y at le a s t understood as

referring to unintentional downward motion. They are all fa irly
fam iliar. Three o f them, ' t u b a r o ' (overturn), 's o rja z o m e ' (collapse)
and ' k u t r u v a l a o ' (roll down) are low r e g i s t e r .

The two best examples r e f e r t y p i c a l l y to CP r a t h e r than CL and are

t h e r e f o r e more l i k e l y to appear i n p r e d i c a t i o n s c h a r a c t e r i z e d i n 2 .4
as 'punctual o c cu r r e n c e s ' . The motion is also vertical and no
'turning' is i n vo l v e d . These p r o p e r t i e s ' sorjazom e' and ' g r e m i z o m e '
share w it h the inclu sive category name, by which they are most
r e a d ily replaceable.

Not ice t h a t 'tu b a ro ' (overturn) i nvolves ' t u r n i n g ' but r a r e l y CL. In
this sense i t i s h a l f - w a y between the two best examples and the two
worst ones, which imply some d u r a t i o n and are typically CL verbs.
For t h i s reason the two worst i nstances o f 'p e fto k a to ' are in f a c t
onl y r e p l a c e a b l e by the hypothesized inclusive category when i t is
used in the environment 'p e fto (kato) apo NP{_oc' (fall (down) from

NpL0c)-

The r e s u l t s of this last set are t h e r e f o r e e x p l a i n a b l e in a fa irly

s t r a i g h t f o r w a r d way in terms o f r e l a t i v e s al i e n c e o f a t t r i b u t e s and
the r e l a t e d noti on o f s u b s t i t u t a b i l i t y (as i t i s used h e r e ) .

Three main poi nts need to be discussed in r e l a t i o n t o p r o t o t y p i c a l i t y

test results. First of a ll, it is quite e v i de n t that subjects
consistently judge some members of verb c a t eg or i e s as more
representative of the category than o th e rs . There can be l i t t l e
doubt t h a t the pr ot ot ype e f f e c t holds f o r the verbs t e s t e d . Similar
evidence i s also provided by Pulman (1983) f o r d i f f e r e n t English verb
categories.

- 312 -
The second p oi n t was a lr ea dy mentioned in preceding s ec t io ns ; it
concerns the p r i n c i p l e ( s ) r e spons ibl e f o r the for mati on o f prototypes
in verb domains. The p r i n c i p l e o f ' f a m i l y resembl ance', which Rosch
and her col leagues have shown t o be r esponsi bl e f o r the formati on o f
prototypes in the domains they have investigated, cannot even be
checked i n the case o f the sets of verbs t e s t e d , at le a s t following
Rosch's method. Reasons f o r this are o f f e r e d here in 1.3 .2 , 4.1.1
and 4 . 1 . 2 . Pulman (1983:111-22) has shown e x p e r i m e n t a l l y that the
'fam ily resemblance' principle is not a causal factor in the
formati on o f protot ypes o f ver bs, even i f a t t r i b u t e s are provided and
analysed by the i n v e s t i g a t o r ( s in ce they cannot be d i r e c t l y obtained
by s u b j e c t s ) .

Pulman makes a number of correct observations concerning the

differences between noun and verb attributes, noun and verb
taxonomies, and how p r o t o t y p i c a l i t y tests i n vo l v i n g verbs are not
amenable t o the same s o r t o f a n al y s i s as are items belonging to noun
domains. S u r p r i s i n g l y enough what seems t o me a most c r u c i a l po i nt
is, however, l e f t o u t . Unl i ke t he t e s t s o f Rosch and her col le agues ,
which u s u a l l y i n vo lv e basic l e v e l and super or di nate c a t e g o r i e s , the
verb taxonomies t e s t e d (both Pulman's and mine) i n v ol ve basic l e v e l
and subordi nate categories.8 At such a low level (subordinate),
differences are not h a l f as e a s i l y d i s c e r n i b l e as a t hi gher l e v e l s .
This e xpl ai ns why "the members o f c at egor i es in question have few i f
any f e a t u r e s in common o t h e r than those which they a l l i n h e r i t from
the c on ta i ni ng cat egor y i t s e l f " (Pulman i b i d . : 1 2 1 ) .

The r e l a t i v e s a l i e n c e o f combinations o f a t t r i b u t e s is shown i n the

SST results to play an i mportant r o l e in s u b j e c t s ' judgments when
they are asked t o 'count' semantic d i s t a n c e . Notice, however, that
t he items used f o r the SST re por te d in the previous sec ti on are
probably ' b a s i c lev e l'. They are at l e a s t very common words in every
day use, perfectly fam iliar, frequent and l i n g u i s t i c a l l y unmarked.
Discussing in terms o f shared p r o p e r t i e s is, in t h e i r case, fa irly
legitim ate. On the o t he r hand, in the p r o t o t y p i c a l i t y t e s t s on the
verbs mentioned here, in or der to secure hyponymic relationships,
subordinate items o f the above ' b a s i c l e v e l ' c at e g o r i e s are i nvol ve d.
This i mpl ie s t h a t a number o f these items are l i n g u i s t i c a l l y marked,

- 313 -
relatively unfam iliar, rather infrequent and, in some cases, only
partially i ncluded i n the hypothesized i n c l u s i v e term. It would be
at least unrealistic to expect such f a c t o r s not t o pl ay a role in
subj ects judgments.

The above remarks l ead to the final and most i mportant p oi n t

concerning the a p p r e c i a t i o n o f the p r o t o t y p i c a l i t y t e s t r e s u l t s , i.e.
what alternatives to the 'family resemblance' principle may be
r espons ibl e f o r the formati on o f prototypes i n verb domains. Rosch
and Mervis (1 975 :5 59) consider the p o s s i b i l i t y o f ' f r e q u e n c y ' pl ayi ng
a r o l e and the p e r c e p t u a l , soci al or memorial salience of p a r t i c u l a r
attributes or particular members of the categories. Pulman
(1 9 83 : 12 0- 3 6) observes t h a t i n the absence o f an independent account
of salience, saying t h a t p r o t o t y p i c a l c a t eg or i e s are the most s a l i e n t
ones is t a ut ol ogous . He also proves experimentally that the
prototypicality judgments he obtained correlate d with semantic
s i m i l a r i t y judgments.

There i s l i t t l e doubt t h a t subj ects a c t u a l l y count semantic di stance

o f hyponyms from the inclusive term and perhaps between co-hyponyms
as w e l l , in view also o f the remarks o f subj ects who c a r r i e d out the
p i l o t t e s t r epor te d here. Not ice also t h a t Rosch and Mervis (1975)
observe t h a t prototypicality ratings p r e d i c t the e x t e n t t o which a
member can be s u b s t i t u t e d for the category name in a sentence. In
short, semantic sim ilarity is, in the case o f hyponymic relations,
almost d i r e c t l y interpretable i n terms o f r e l a t i v e s u b s t i t u t a b i l i t y .
There i s a l o t o f evidence f o r t h i s in the t e s t s r epor te d here, both
concerning c at e g o r i e s which y i e l d very r e l i a b l e results, i.e. great
inter-subject agreement on the p o s i t i o n o f 'genuine hyponyms' which
are s u f f i c i e n t l y removed from one another and from the s uper or di nate
- (e .g. category A ) , as well as c at e go r ie s yielding l e ss reliable
results (e.g. categories G and H). The l a t t e r categories i nv ol ve
'dubious cases' which are called here i nstances of 'p artial
inclusion' and o f f e r poor s u b s t i t u t a b i l i t y p o s s i b i l i t i e s . 9

Notice, further, th a t within a taxonomy one does not need to s p el l

out all the properties shared, but simply to identify poi nts of

- 314 -
d e par t ur e {dissim ilarity from the category name). In the case of
re la te d low-level categories (subordinates), such a t a sk i s c e r t a i n l y
more f e a s i b l e than counting number o f attributes. I t is also more
realistic and does not presuppose t h a t each i n d i v i d u a l attribute is
recognized as such, i.e. i t i s e q u a l l y compatible with a ' h o l i s t i c '
as wel l as w i t h a ' c om p o n e n t i a l ' understanding o f c a t e g o r i e s . This
is fa irly easily done in the case of sets including items all of
which are almost e q u a l l y l i n g u i s t i c a l l y unmarked and f a m i l i a r . Take
as an example cat egory A, i.e. 'p e r p a t a o ' (wal k) and i t s hyponyms.
'P er ce p tu a l salience' of attributes and c at e g o r i e s can be readily
interpreted in terms o f semantic s i m i l a r i t y and s u b s t i t u t a b i l i t y in
this case. If some a ct of 'walking' is somehow 'defective', we
expect i t to be judged as an a t y p i c a l act o f 'walking' and t h i s is
c l e a r l y the case w i t h ' t r i k l i z o ' ( s t a g ge r ) and ' p a r a p a t a o ' ( s tu mb l e) .

If, however, semantic s i m i l a r i t y and notions r e l a t e d to i t were the

onl y f a c t o r s r e s po ns i bl e f o r the formati on o f protot ypes in the case
o f verbs, a number o f r a t i n g s would have been d i f f e r e n t , e . g . :

(a) 'fu d a ro ' (sink to the bottom) would never be judged the
worst i nstance o f ' v i B i z o m e ' ( s i n k ) in c at egor y B.

(b) 'ip e rp iS a o ' (jump over) would never appear as the worst
example o f 'p /5ao' (jump) in category C.

(c) 'a x B o fo ro ' (carry baggage) could receive a sim ilar rating
to 'k u v a l a o ' (carry), i.e. appear in the 2nd r a t h e r than
the pe n u l t i ma t e position within set D ('p/yeno^' and its
hyponyms).

(d) 'p r o e l a v n o ' (advance (m ilit.)) would be much closer to

'p r o x o r o ' (advance) in category E ( i n c l u d i n g 'p/yenoj' (go)
and i t s hyponyms).

( e) ' ksekubizom e' (clear off) would not have taken the 5th
position in category F, i.e. 'fe v y o ' (leave), as it

- 315 -
 certainly has fewer 'extras' by comparison wi th the
i n c l u s i v e cat egory than the items t a k i n g up the 1 s t , 2nd and
3rd p o s i t i o n s .

In s h o r t , ot he r f a c t o r s are also i nvol ve d, the most obvious of which

have a lr e ad y been mentioned, namely fam iliarity, frequency and
linguistic markedness, which are not completely distinct from one
another, but do not amount t o the same t h i n g e i t h e r , as I have t r i e d
to demonstrate i n the p r e s e n t a t i o n o f the t e s t r e s u l t s .

Not ice firs t, that high or low register items may, however, be
unfamiliar, i n f r e q u e n t and/ or l i n g u i s t i c a l l y marked. This s i t u a t i o n
is l ess l i k e l y to a r i s e in the case o f items which do not belong to a
marked register, for the obvious reason that so ciolinguistically
marked words are u s u a l l y o f more r e s t r i c t e d use, i.e. l ess f r e q u e n t .
Ne ver the le ss , frequency and fam iliarity do not c oi n ci d e entirely
either. Consider, for i ns t an c e, the best examples of category E,
namely 'p o re vo m e ' (go a long di stance on fo ot) and 'p r o x o r o '
(advance). They are equally well-known to speakers, i.e. quite
fam iliar, but the f i r s t one is c e r t a i n l y much less f re q u e n t than the
second. Wi thi n t he same cat egor y, the worst examples, 'p a re la v n o '

(march) and 'p r o e l a v n o ' (advance (mil i t . ) ) are probably equally

infrequent, but the former one i s c e r t a i n l y much b e t t e r known (more
f a m i l i a r ) than the l a t t e r .

F a m i l i a r i t y and frequency, understood in the suggested way, are shown

to pl ay a r o l e in p r o t o t y p i c a l i t y judgments. They are also d i s t i n c t
to some e x t e n t from l i n g u i s t i c markedness, which also plays a r o l e in
such judgments. Consider the verb ' ksekubizom e' (clear o f f ) , which
is certainly fa m ilia r to the subjects and fa irly f re que nt but
l i n g u i s t i c a l l y marked. It is suggested in the di scussion o f category
F t h a t t h i s f a c t o r ( l i n g u i s t i c markedness) i s r e pons ibl e f o r the f a c t
that the item i n question is not judged a p r o t o t y p i c a l i nstance o f
'fe v y o ' ( l e a v e ) w i t h which i t covers the same conceptual a rea .

No t i c e , f i n a l l y , t h a t i f only f a m i l i a r i t y was the d e c i s i v e f a c t o r f o r

such judgments, a number o f r a t i n g s would have been d i f f e r e n t , e.g .:

- 316 -
 ( a) 's a lta ro ' (jump, leap) would never be judged the best
i nstance o f 'p ib a o ' (j ump).

(b) 'p a r a p a t a o ' (stumble) would never be judged almost the worst
i nstance o f 'p e rp a ta o ' (walk).

( c) 'ka ta p o n d iz o m e ' ( s i n k t o the bottom) would not be judged the

2nd best example o f ' v i Q i z o m e ' (sink).

(d) 'ta k s ib e v o ' (travel) would be judged c l o s e r t o t he category

name ('p iy e n o j' ( go) ) than 'porevom e' (go a long way on
foot). The exact opposite is a c t u a l l y the case.

(e) 'b ra p e te v o ' (escape) and even 'm e ta n a ste vo ' (e mi gr at e )

within the Greek s oc ia l context are more fam iliar than
'a n a x o ro ' ( d e p a r t ) and ' a p o x o r o ' (withdraw) and would not be
judged the l e a s t t y p i c a l o f the i n c l u s i v e c at egor y o f which
the l a t t e r p a i r o f verbs are judged as the best examples.

The picture which emerges concerning the interplay of the factors

discussed and e x e m p l i f i e d above has t he f o l l o w i n g c h a r a c t e r i s t i c s :

Hyponyms are ranged along d i f f e r e n t dimensions and each category has

c h a r a c t e r i s t i c s speci al t o it. To q u a l i f y as a p r o t o t y p i c a l i nstance
o f a c at egor y, a member must be e a s i l y r e p la c e a b l e by the inclusive
category name, and convey a s i m i l a r p i c t u r e to the one conveyed by
the i n c l u s i v e c at egor y (i.e. not be 'defective' in any way). This
may be also expl ai ned in terms of attributes; prototypical members
differ from t he i n c l u s i v e category a t few and r e l a t i v e l y unimportant
points. In addition to t h i s , prototypical members are normally
fa irly fam iliar, f re q u e n t and lin g u istically unmarked. The latte r
set of properties seems t o p l a y a less d e c i s i v e role than semantic
di s ta n ce repr esented in terms o f substitutability and r e l a t i v e
salience of p roperties.

- 317 -
 Notes on Chapter 5

1. I am r e f e r r i n g mainly to Fillenbaum and Rapoport (1971), M iller

(1969, 1971, 1 972), Long ( 1 9 7 5 ) , Rosch and Mervis (1975), Rosch
et a l . (1976), Pulman 1983).

2. This does not e x p l a i n why 'b ia s x iz o ' (traverse) i s also l i n k e d to

the 'random wal k' verbs. It must be a p p r e c i a t e d , though, that
while forming cluste rs subj ects of t en group two items t o g et he r
(e.g. 'p e rn ao ' and ' b i a s x i z o ' ) as most c l o s e l y r e l a t e d in meaning
and then add o t h e r items t o the ' nu c le us ' o f t he c l u s t e r , which
may be l i n k e d w i t h only one o f the two f i r s t members. In s h or t ,
they are aware o f certain attributes as connected w i t h certain
items and may then over -extend their applicability to cover a
whole set, although the attribute(s) in question may not be
present in t he whole s e t . This may also be the case w i t h c l u s t e r
C and the inclusion of ' sernome' ( cr eep, crawl), a lr ea dy
discussed. There is some evidence that this procedure was
f o l l o w e d , from those l i s t i n g s o f items which were w r i t t e n down on
paper (by the s u b je ct who had formed the corresponding c l u s t e r s
and decided t o 'ease' my t a s k ) . Some obvious p a i r s come f i r s t ,
and these are f ol l ow ed by l e ss c l e a r l y r e l a t e d i tems.

3. In view o f the f a c t t h a t 'p/yeno,?' ( t a k e t o ) and ' f e r n o ' (bring)

c o - e x i s t also in a compound word 'p i y e n o f e r n o ' ( t a k e and b r i n g ) ,
one could expect some l i s t i n g s o f ' f e r n o ' with 'p iy e n o '. This is
never the case, because c r o s s - c l a s s i f i c a t i o n i s not allowed and
'p iy e n o ' i s always i n t e r p r e t e d as ' g o ' , as a lr e a d y noted. Notice
also t h a t 'fe rn o ' (bring) differs from 's te ln o ' (send) not only
in terms o f presense vs absence o f 'accompaniment' but also in
the sense t h a t only the former i mpl ie s t h a t the o b j e c t caused to
move reaches i t s d e s t i n a t i o n . The l a t t e r verb i s analysed in 3 . 2
as implying onl y t h a t the o b j e c t i s caused t o ' s t a r t moving away
from the causer'. Despit e these differences, which are also
present in the equivalent English verbs, ' send' is o ft e n used

- 318 -
 parallel to 't a k e ' as the c au sa ti ve c o u n te r pa rt o f ' g o' and t h e i r
difference is reflected in calling 'send' an i nstance of
'b a llis tic ' and 'bring' an i nstance of 'controlled' causation
( e . g . C l a r k 1 9 7 4: 3 22 ) .

4. Rosch's i n s t r u c t i o n s are u s u a l l y o f the form: "Rate on a 7 - p o i n t

scal e the extent to which instances of common s uper or di nate
c at e gor i es r e p r es e nt your ' i d e a ' or 'image' o f the meaning o f the
category name". The equivalent MG paraphrase seemed to me
extremely e llip tic to be used wi th Greek s tude nts , who are
accustomed t o more e x p l i c i t i n s t r u c t i o n s .

5. Since ' e x p ec t e d' here is 'random' ( t e s t i n g the n u l l h y p ot he s is ) ,

the more removed the ' observed' result is from the 'expected',
the s tr onge r the assumption t h a t the d i s t r i b u t i o n i s not random.
In short, great inter-subject agreement is repr esented by a
number (much) h i ghe r than 1 2. 5 9. Anything l ower than 12.59
i mpl ies not reaching s i g n i f i c a n c e .

6. The i n c l u s i v e term here is ' p / y en o^ ' understood as a causati ve

and marked as such in t he test-booklets ('take something
somewhere'), as the 'semanti c sim ilarity' test showed that
although ext remel y common t h i s understanding i s secondary to the
non-causati ve 'p /y e n o j' (go).

7. The d i f f e r e n c e i n l i n g u i s t i c markedness i s s l i g h t , because except

for the 'affective ov er tone' of 'un d e s ira b ility', the verb is
otherwise ext reme ly common, f a m i l i a r and o f ever y-day use.

8. A l a r g e number o f Rosch's t e s t s on p r o t o t y p i c a l i t y are r eported

in Rosch and Mervis (1975:573-605). The types examined i nvolve
only super or di na te c at e g o r i e s ( f u r n i t u r e , vehicle, fru it, weapon,
v eg e ta b le , clothing) and ' b a s i c l e v e l ' items which are hyponyms
of the c a t e g o r i e s j u s t mentioned. Al l the experiments r eported
ask subj ects for listings of attributes o f such c a t e g o r i e s , the
'lowest' level ones being: car, truck, airplane, chair, table,
i.e. basic l e v e l , common concrete nouns.

- 319 -
 In Rosch e t a l . ( 1 9 7 6 :3 8 2- 4 39 ) where the investigation concerns
identifying the 'basic level of abstraction', and not
prototypicality, o b j e c t names are given (as v a r i a b l e s ) belonging
to all three levels of abstraction (i.e . i n c lu d i n g also
subordinates). Notice, however, that only two subordinates are
provided f o r each hypothesized basic l e v e l cat egor y, e.g. 'flo o r
lamp' and 'dress lamp' for 'lamp', 'c ity bus', 'cross country
bus' f o r ' b u s ' .

9. Pulman ( 1 983 :1 35) claims that category membership (and class

inclusion) i s an a l l or nothing m a tt e r and provi des examples in
support o f t h i s c la i m . N o t i c e , however, t h a t Pulman focuses on 8
c a t e g o r i e s drawn from the whole verbal vocabul ary o f English and
comes up w i t h sets including only genuine hyponyms. 'Relative
class inclusion' is discussed here, because 'semi-hyponyms' had
to be included i f a s u f f i c i e n t number o f sets was t o be t e s t e d .
I t seems to be the case t h a t 'bad' examples o f hyponymic sets are
almost equally informative wi th 'good' ones. A unified
e x pl a na t io n can be based on both kinds o f s e t s.

- 320 -
 CONCLUSIONS

Motion verbs have been extensively analysed within the framework of

traditional semantic theories such as semantic field theor y and
componential analysis. They constitute a well-established semantic
field , highly structured and exhibiting a number of properties also
relevant to o the r verb domaims. MG verbs o f motion, which have never
previously been i n v e s t i g a t e d , exhibit certain idiosyncrasies, owing to
the aspectual system o f Modern Greek and t o sociolinguistic factors.
Ne ver the le ss , they r e f l e c t on the whole a semantic s t r u c t u r e s i m i l a r t o
t h a t o f most Indo-European languages.

Prototype t he o ry was adopted as the most suitable method of

i n v e s t i g a t i o n o f t h i s domain, because i t r e s t s on some basi c assumptions
on word meaning which are i n t u i t i v e l y more convincing than those of
traditional semantic t h e o r i e s . The most i mportant o f these assumptions
is that word meaning i s not a m a t te r of necessary and sufficient
c o ndi ti ons but a graded phenomenon, and that therefore, semantic
c a t eg or i e s have b l u r r y edges and a ll o w f o r degrees o f membership. The
v a l i d i t y o f such a working hypothesis needs to be demonstrated by being
t e s t e d i n as many d i f f e r e n t areas as p o s s i b l e , as well as by having i t s
r e s u l t s compared a g ai n st those o f a l t e r n a t i v e methods o f i n v e s t i g a t i o n .

In examining analyses o f MVs carried out within t he framework of

componential analysis and semantic field th eories, it became obvious
t h a t one o f t h e i r most i mportant shortcomings was the o b l i g a t i o n which
they imposed to c o n st r u c t symmetrical tables and/ or neat formulae
utilizin g the smallest number of fe atures. Such economical and/ or
contrast based solutions were shown to result in definitions which
i nvolved p r o p e r t i e s not necessary f o r the s p e c i f i c a t i o n o f c e r t a i n items
and l e f t out i mportant i n f or m at io n which was not amenable t o t he desi red
form alization. I t was demonstrated t h a t d e s c r i p t i o n s o f MVs c a r r i e d out
within the framework of alternative lexical semantic theories were
l a c k i n g i n d e s c r i p t i v e adequacy.

- 321 -
In the course o f the a n al y s i s o f MGMVs, a distinction was e st a b l i s h e d
between ' maj or classificatory properties' and 'mi nor properties'
r e l e v a n t to the s p e c i f i c f i e l d under i n v e s t i g a t i o n . The former type of
p r o p e r t i e s was analysed f i r s t . The c a t e g o r i a l frameworks which i nvolve
the notions o f ' s t a t e s - p r o c e s s e s - e v e n t s ' , ' c a u s a t i v i t y ' and ' a g e n t i v i t y '
were understood as r e l e v a n t t o a c a t e g o r i z a t i o n o f many d i f f e r e n t areas
of the verbal vocabul ar y, related to grammatical categories such as
'aspect' and ' t r a n s i t i v i t y ' , a p p l i c a b l e to a characterization o f whole
predications - and individual verbs through predications - and
anal ys abl e into clusters of scalar properties. I t was also shown t h a t
such properties are on the whole paradigmatically related to one
another.

The 'states-processes-events' (S-P-E) d i s t i n c t i o n was examined in some

detail in Chapter 2. I t was demonstrated t h a t S-P-E i s best understood
as a continuum, the f oc al poi nt s o f which are i d e n t i f i a b l e on the basis
of the interaction of a number of factors such as 'aspect', spatio-
temporal specifications, i n h e r en t verb p r o p e r t i e s such as ' d u r a t i o n ' and
q ualities of the 'theme' such as 'count' vs 'mass' and singular vs
plural. On t he basis of such consi de ra ti ons it has been p os si bl e to
i d e n t i f y what c o n s t i t u t e s c e n t r a l instances o f a ' s t a t e ' a 'process', a
'development' and a ' punctual occurrence' within the field under
i n v e s t i g a t i o n and c o n st r uc t ' t e s t - f r a m e s ' in or der to check t he r e l a t i v e
degree o f ' p r o c e s s u a l i t y ' and ' e v e n t - l i k e n a t u r e ' of a number o f MGMVs.
S i m i l a r d i s t i n c t i o n s are expected in ot he r languages where the continuum
i s probably cut a t poi nts identifiable wi th the help o f d i f f e r e n t , to
some e x t e n t , l i n g u i s t i c means.

Causativity and a g e n t i v i t y were discussed in Chapter 3 where it was

suggested that the relevant facts pertaining to them could be b e t t e r
approached i f the two notions were kept d i s t i n c t , to some e x t e n t , and
t he a n al y s i s o f t he former concentrated on the q u a l i t i e s of the caused
event, while the investigation of the latte r focused on t he type of
causing e vent. Both noti ons were also understood as i n v o l v i n g conti nua.

The degree o f c a u s a t i v i t y o f t r a n s i t i v e verbs was shown t o depend mainly

on the degree of deviation from the protot ype of t he patient. The

- 322 -
degree of a g e n t i v i t y o f both t r a n s i t i v e and i n t r a n s i t i v e verbs was shown
to depend mainl y on the degree o f d e v i a t i o n from the prot ot ype o f the
agent. It was demonstrated that factors posi ted by alternative
approaches as c r u c i a l f o r the i d e n t i f i c a t i o n o f c a us a t i ve verbs such as
1e x i c a l i z a t i o n of the resulting condition/position of the causee were
not significant on the basis o f the data examined. It was also shown
that in the case of phonologically identical causative/non-causative
p a i r s o f MG Vs, t he non-c aus ati ve member is more ba si c.

The distinction between direct/indirect causati ves was considered

i nadequate. Both w i t h i n the m an ip u l a t i v e and the non-ma ni pul at ive type
o f causati on d i f f e r e n t grades could be i d e n t i f i e d and t he phenomenon was
analysed in terms of a line l e adi ng from e xp licit causati ves to
prototypical instances o f d i r e c t c aus at ion. The degree o f affe ctednes s
of the object of MG transitive const ruct io ns was related to
passivizability; the borders between MG '-o m e ' intransitives and
passives were shown t o be i n d et e rm i n at e ; passives were treated along
w it h i n t r a n s i t i v e s , and i t was shown t h a t even ' g e n u i n e l y ' pa ssi ve-f orm
verbs can be more a gent i ve than c e r t a i n a c t i v e - f o r m i n t r a n s i t i v e s .

The r e l a t i o n between the MVs under examination and t h e i r p r o p e r t i e s were

discussed in Chapter 4. I t was f i r s t demonstrated t h a t verb taxonomies
d iffer s ig nificantly from noun taxonomies of the t ype extensively
analysed in e t h n o s c i e n t i f i c s t u d i e s . In most cases, onl y two l e v e l s are
safely identifiable, ge ner i c and specific; a number of taxons
participate i n competing taxonomies; many instances o f s u b s t i t u t a b i l i t y
are a t t r i b u t a b l e t o p a r t i a l r a t h e r than proper class i n c l u s i o n , so t h a t
w i t h i n any s p e c i f i c verb domain, few c l e a r cases o f genuine hyponyms can
be i d e n t i f i e d . Besides, since verbs i nv ol ve on the whole more a b s t r a c t
and complex categorial frameworks than most types of nouns, it is
virtu ally impossible to obtai n listing s of attributes directly from
n a t i v e speakers.

Owing to the above arguments, Rosch's method o f identifying the basic

level of abstraction is inapplicable in the case of verbs and an
a l t e r n a t i v e method was t e n t a t i v e l y proposed which i nvolved t a k i n g into
c o ns id e r a t i o n factors such as re lative substitutability (and class

- 323 -
inclusion), l i n g u i s t i c markedness, frequency and f a m i l i a r i t y . The same
factors, as wel l as the r e l a t i v e s al i e n c e o f a t t r i b u t e s were considered
p e r t i n e n t t o the for mati on o f protot ypes of verb c a t e g o r i e s . Wi thi n the
area o f MGMVs, 'minor' properties (category a t t r i b u t e s ) were i d e n t i f i e d
and attention was drawn to the interdependence of 'manner',
'instrum entality' and 'medium', and the i n t e r r e l a t i o n s between 'manner'
and 'd ire c tio n a lity '. Besides, a number of ot he r combinations of
properties was shown to be non-arbitrary such as 'inward' wi th
'downward' or 'downward' w it h ' u n i n t e n t i o n a l ' .

In or de r t o t e s t the v a l i d i t y o f some o f the above observati ons agai nst

the intuitions of native speakers, as wel l as to check whether the
principle of t he non-equi val ence of category members and the
p r o t o t y p i c a l i t y e f f e c t obtained f o r verbs, a semantic s i m i l a r i t y s o r ti n g
t a sk (SST) and a p rototypicality test were carried out involving MG
verbs o f motion.

The SST r e s u l t s ( r e p o r t e d i n Chapter 5) provided some evidence in favour

o f the p r o t o t y p i c a l r a t h e r than the 'most g e n e r a l ' understanding o f verb
categories, and the salience of specific combinations of properties
(such as v e r t i c a l i t y and c a u s a t i v i t y ) , which is shown t o c o n s t i t u t e a
more p l a u s i b l e e x pl a n a t i o n o f semantic s i m i l a r i t y judgments than number
of shared attributes and h i er a r c h y of individual and independent
properties.

The p rototypicality test results showed that the principle of non­

equi val ence o f cat egory members was v a l i d f o r the verbs t e s t e d , d e spi t e
the horrendous c ompl icati ons inherent in the material investigated
(owing to the peculiar sociolinguistic situation of MG). It was
demonstrated t h a t t he items c o n s i s t e n t l y judged as the most p r o t o t y p i c a l
instances o f t h e i r i n c l u s i v e cat egory were the ones which could be most
e a s i l y r eplaced by t he cat egory name and which conveyed the most s i m i l a r
picture to it, i.e. had the s m al l e s t semantic d i s t an c e from their
s u p er o rd i na t e . Semantic s i m i l a r i t y being a lr eady equated w i t h r e l a t i v e
s a l i e n c e o f combinations o f p r o p e r t i e s , the l a t t e r emerges as the most
crucial factor in t he f or mati on of prototypes in the area under

- 324 -
investigation. Linguistic markedness, frequency and fa m iliarity were
also shown to pl ay an i mpor ta nt , though less d e c i s i v e , role.

The r e l a t i o n between a t t r i b u t e c l u s t e r s and the formati on o f c at e gor i es

remains a most interesting i ss ue , in need of further research.
Additional e mp i r i c a l corroboration for the existing hypotheses on
lexical meaning needs to be accumulated and p r o p e r l y e val uated before
such hypotheses can be r ef or mu l at e d and incor por at ed w i t h i n a model of
grammatical d e s c r i p t i o n which can c lai m psychological r e a l i t y .

- 325 -
 APPENDIX

LIST I ; Modern Greek verbs o f motion and p o s i t i o n

A agizi B A touches B
A akoluBi B A follows B
A akoluQjete A i s fol lowed
A akubai2 se C A leans on C
A akubaig B se C A puts B on C
A anakatevi B A stirs B
A anakatevete A i s s t i r r e d , churns
A anapibai A jumps up (and down}, bounces
A anapoboyirizij A o v er t ur ns , turns upside down
A anapoboyirizi% B A overturns B, tur ns B upside down
A ana po bo y ir iz et e A i s turned upside down,
is overturned
A aneveni A ascends, goes up
A anevazi B (se C) A takes B up, r a i s e s B ( t o / o n C)
A anevokateveni A goes up and down
A anevokatevazi B A takes B up and down
A apoyioni B A causes B t o take o f f
A apoyionete A takes o f f
A apomakrini B A takes B away, removes B
A apomakrinete A moves away, is taken away/removed
A apoplei A sails o ff
A armenizi A s a i l s about
A bazi B se C A puts B in C
A beni se C A gets i n / e n t e r s C
A b i z i B se C A s t i c k s B in C
A b i z e t e se C A is stuck/sticks i t s e l f in C
A busulizi A crawls (as o f a baby)
A bja ven i ( C) A passes through/crosses (C)
A b r a s k e l i z i (C) A strides (over C)
A e k s fe nb o ni z i B A hurls/flin g s/slin g s B
A eksfenbonizete A throws i t s e l f , i s hurl ed
A ekt oks evi B A launches B

- 326 -
A ektoksevete A i s launched
A epiplei A floats
A e pistrefi A comes back, r e t u r n s
A erxete A comes
A f e r n i B (se C) A brings B ( t o C)
A fevyi A goes away, l e a v es , departs
A f t a m ' i ( se C) A arrives at/reaches C
A f t a m '2 B se C A causes B t o reach C
A gr em iz i B A p u l l s B down (a p r e c i p i c e )
A gremizete A f a l l s down/is caused to f a l l down
(a p r e c i p i c e ) , col la ps es
A y lis tra ij A slides, slips
A ylistra i'2 B A causes B to s l i d e
A y iriz ii A goes around, tu rn s
A y irizi'2 B A tur ns B
A \1irizi3 ( e p i s t r e f i ) A goes back ( r e t u r n s )
A y i r i z i / j . B se C A r e t u r ns B t o C
A yona tizij A kneels
A \0 natizi2 B A causes B t o kneel
A ipoxori A gives way, goes down under
pressure, recedes, subsides
A ip s o n i B A raises B
A ipsonete A rises, i s r ai se d
A kade tej (se C ) / i n e ka ismenos A i s seated (on C)
A kaBete2 se B A s i t s / i s s i t t i n g on B
(assumes a s i t t i n g p o s i t i o n )
A kaQizi B A causes B t o s i t
A kalpazi A gal lops
A katabioki B A chases B
A kat aftiokete A i s chased/pursued
A kateve ni A goes down, descends
A k a t e v a z i B (se C) A takes B down ( t o / o n C)
A ka tra kila i A roll down
A k ik lo fo rij A c i r c u l a t e s , goes around
A k ik lo fo n '2 B A circulates B
A k ila ij A rolls
A k i l a i ' 2 B (se C) A causes B t o r o l l ( t o C)
A k ilje te A wallows

- 327 -
A k in iya i B A hunts/ chases/ runs a f t e r B
A koliba i A swims
A kremai B se C A hangs B on C
A kremete (se C ) / i n e kremasmenos A is hangi ng/ hangs /i s hung (on C)
A kremjete A hangs i t s e l f , i s hung
A k s a n a y i r i z i (se C) A goes/comes b a ck / r e t u r n s ( t o C)
A ksa n a (e )r xe te (se C) A comes back ( t o C)
A k s ap lo m 'j (se C ) / A lies /is l y i n g down (on C)
ine ksaplomenos
A ksaplom'2 (se C) A is l y i n g down (on C)
(assumes a l y i n g p o s i t i o n )
A k s a p l o n i ^ B (se C) A causes B t o l i e down (on C)
A ksekinai A starts o ff
A k u n j e t e ( k u n a i2) A moves, s t i r s
A k un a i2 B A causes B t o move, shakes B
A kutruvalai A r o l l s down
A kuvalai B A carries B
A kuvaljete A is carried
A metaferi B A transports B
A me taferete A i s tr an s po rt e d
A metakomizi1 A moves house
A metakomizi2 B A moves B ( f u r n i t u r e )
A ine meteoros A is (suspended) in the a i r ,
i s dangling
A obi y i B A leads B
A o biyite A i s led
A pa r a m e r iz i 2 A moves a s i d e / o v e r
A p a r a m e r iz i2 B A puts B aside
A paramerizete A is put aside
A p a r a p a ta i A stumbles
A p e f t i (se C) A falls (on C)
A p e rife ri B A c a r r i e s B about, causes B to
go about
A perife re te A roams around, i s c a r r i e d about
A p e rik ilo n i B A encircles B
A perikiklo nete A is e n c i r c l e d
A perip ian jete A wanders
A pernaij A passes

- 328 -
A p e r n a iz B A cause's B to pass
A p e r p a ta i A walks
A petaij A flies
A p etai2 B A throws B
A pe tje te j/p e ta ye te j A i s thrown (away)
A p e t j e t e 2 / p e t a y e t e 2 (se C) A dashes/goes f o r a short time
( t o C ) , jumps up
A p i y e n i j se C A goes t o C
A piyem'2 B se C A takes B to C
A pi yen oerxete A comes and goes
A piyenoferni B A takes B s.wh. and bri ngs i t back
A p i b a i (B) A jumps (over B)
A p lei A sails
A p lis ia z ij B A approaches B
A p l i s i a z i ' 2 B se C A causes B to approach C
A porevete A goes/walks (a long d i st a n c e )
A prosyioni B A lands B
A pr o s y io n e te A lands
A pr o s p e rn a i2 B A overtakes/ passes by B
A rix n i B A throws B
A r i x n e t e (se C) A throws i t s e l f / i s thrown t o / f a l l s
(on C)
A salevi A moves s l i g h t l y / s t i r s
A se rja n izi A walks around
A semi B A drags B
A sernete A drags, i s dragged, creeps, crawls
A s ii B A shakes B
A siete A shakes
A sikoni B A ra is e s /lifts B
A siko net e A r i s e s , gets up, i s l i f t e d
A sirtobevi B A accompanies B
A sinobevete A i s accompanied
A s k a r f a l o n i (B) A climbs (up B)
A skivi A bends
(A sproxni B) A pushes B
A stazi A d r ips
A steki/stekete A stands

- 329 -
A s tin i B A causes B to stand
A stinete A i s caused to stand, assumes a
standing p o s i t i o n
A strivi'2 (B) A t urns (around B)
A s triv i'2 B A tur ns B
A s trifo y iriz ij A t w i s t s and turns
A s trifo y iriz i2 B A causes B to t w i s t and t ur n
A sulatsari A strolls, saunters
A t a k s iS e v i A travels
A talandevi B A causes B t o o s c i l l a t e / s w i n g
A taland eve te A oscillates, swings
A t a ra ku na i B A shakes/jolts B
A tarakunjete A shakes, is shaken/jolted
A tarazi B A shakes/agitates B
A t a ra z e te A shakes, is shaken/stirred
A tin a z i B A shakes B up, throws B a b ru p t l y
A tinazete A shakes/ jumps/spri ngs up, j e r k s ,
i s shaken up
A tradazi B A shakes B
A t r a d az et e A shakes, is shaken/jolted
(A t r a v a i B) A pulls B
A tremi A trembles
A trexi 2 A runs
A t r e x i '2 B A causes B t o run
A triy irn a ij/triy iriz ij A roams around, goes here and t h er e
A triy irn a i2 B / t r i y ir iz i2 A causes B to go here and t he re
A tsu la i 2 A ro lls, slides
A t s u la i '2 B A causes B to r o l l / s l i d e
A vazi B se C A puts B i n / o n / a t C
A vyazi B (apo C) A takes B out ( o f C)
A vyeni (apo C) A goes out ( o f C)
A viB izi B A sinks B
A viBizete A s in k s, i s sunk
A vo lta ri A walks about
A v r i s k e t e se C A i s (found) i n / o n / a t C
A vul ja zi '2 A sinks
A vu lja zi2 B A sinks B
A v u t a i 2 B se C A immerses/dips B i n C

- 330 -
A v u t a i j ( k an i v u t j a ) A dives
A v u t j e t e se C A gets immersed i n t o C
A xam ilonij A lowers, stoops
A xamilom'2 B A lowers B
A x in i B A pours B
A x i n e t e (se C) A is poured/spilled, flows ( i n t o C)
A xo ni B se C A sticks B i n ( t o ) C
A xonete se C A s t i c k s / i s stuck i n ( t o ) C,
gets s t uc k/ e n g u l f e d i n ( t o ) C
A xoropibai A hops, jumps about
A xorevij A dances

- 331 -
LIST I I : Fi ve t e s t frames f o r Modern Greek verbs o f motion

(a) A (B) ( j a ) 10 metra / pontus / x i l j o m e t r a

A (B) (for) 10 metres / c ent imetr es / k i l om e t r es

akoluBo, akoluBjeme; aneveno, anevazo; apomakrino, apomakrinom;

busulizo; y i r i z o j ^ ; y lis tra o j kalpazo; kat abi oko , katabiokome;
kateva zo, kateveno; k a t r a k i l a o ; k i k l o f o r o ; k i l a o j k i l j e m e ;
k i n i y a o ; k o l i b a o ; k u t r u v a l a o ; kuvalao, kuvaljeme; m e ta fe r o ,
metaferome; obiyo, obiyume; param erizo , paramerizome; perpatao;
p e t a o j ; piy en o; prospernao; v i B i z o , viBizome; v u l j a z o j 2* vuta oj 2

(b) A ________ (B) apo a f t o to simio mexri ekino

A ________ (B) from t h i s p o i n t up to t h a t one

akoluBo, akoluBjeme; aneveno, anevazo; anevokateveno, anevokatevazo;

b u s u l i z o ; eksfenbonizo, eksfenbonizome; ektoksevo, ektoksevome;
erxome; f e r n o ; ( f t a n o ) ; gremizo, gremizome; y i r i z o j 2* y l i s t r a o j 2>
ip oxoro; ipsono, ipsonome; kalpazo; k a t a b i o k o , katabiokome; katevazo,
kateveno; k a t r a k i l a o ; k i l a o j 2 > k i l j e m e ; k o l i b a o ; ksanaerxome;
k s a n a y i r i z o ; k u t r u v a l a o ; k u v a la o , kuvaljeme; metafero, metaferome;
obiyo, obiyume; p e r i k i k l o n o , p e r ik ik lo n o m e ; perpatao; pi yen o; pibao;
p l e o ; p l i s i a z o ; r i x n o , rixnome; (p ara)serno, (para)sernome; sikono,
sikonome; sinobevo, sinobevome; s k a r fa lo n o ; ( spr ox no) ; s t r i v o j 2»
ta k s ib e v o ; t i n a z o , tinazome; t r av a o, travjeme; t r e x o j ^ i t s u l a o j 2:
vazo; v i B i z o , viBizome; v u l j a z o j v u t a o j ^ ; xam il on oj 2* x i n o ,
xinome; xono, xonome

- 332 -
(c) A (B) apo a f t o to simio se ekino
A _______ (B) from th is point to th a t one

akoluBo, akoluBjeme; (akubao2); aneveno, anevazo; anevokateveno,

anevokatevazo; (bazo, beno); b u s u l i z o ; eksfenbonizo, eksfenbonizome;
ektoksevo, ektoksevome; erxome; f e r n o ; f t a n o j ^ S gremizo, gremizome;
Y i r i z o j t 2 33 i Y? i s t r a o j 2* ipo xor o; ipsono, ipsonome; kalpazo;
kat abi oko , katabiokome; katevazo, kateveno; (kaBome, k a B iz o ) ;
katra kila o; k ik lo fo ro ; k i l a o j k i l j e m e ; kiniyao ; kolibao; (kremao,
kremome); ksanaerxome; k s a n a y i r i z o ; ksaplono; k u t r u v a l a o ; kuvalao,
kuvaljeme; meta fero, metaferome; metakomizoj 2* obiyo, obiyume;
parapatao; p e f t o ; p e r i k i k l o n o , p e r ik i k lo n o m e ; pernaoj perpatao;
p e t a ° j }2> petjeme/petayomej 2* piy en o; piyenoerxome; p i y e n o f e r n o ;
p ib a o ; p l e o ; p l i s i a z o ; r i x n o , rixnome; serno, sernome; sikono,
sikonome; sinobevo, sinobevome; s k a r f a lo n o ; ( spr ox no) ; ( s t i n o ,
stin om e) ; s t r i v o j ^ : tak si be vo ; t in a z o , tinazome; ( t r a v a o ) ; t r e x o j £»
t s u l a o j ^ » ( v az o) ; vyazo, vyeno; v i B i z o , viBizome; v u l j a z o ; vut aoj
xam il ono j 2> x i n o , xinome; (xono, xonome); ( x o re v o ); xoropibao

(d) A ________ (B) apo a f t o to simio

A ________ (B) from th is point

akoluBo, akoluBjeme; aneveno, anevazo; anevokateveno, anevokatevazo;

apoyiono, apoyionome; apomakrino, apomakrinome; apopleo; armenizo;
bazo, beno; b u s u l i z o ; eksfenbonizo, eksfenbonizome; ektoksevo,
ektoksevome; erxome; f e r n o ; f e v y o; f t a n o ; gremizo, gremizome;
y i r i z o 3; y l i s t r a o j 2* ip oxoro; ipsono, ipsonome; kal pazo; kat abi oko ,
katabiokome; katevazo, kateveno; k a t r a k i l a o ; k i l a o j 2 > k i l j e m e ;
k i n i y a o ; k o l i b a o ; kremao, kremjeme; ksekinao; kunjeme, kunaoj;
k u t r u v a l a o ; kuvalao, kuvaljeme; metafero, metaferome; metakomizoj 2:
obi yo, obiyume; paramerizo, paramerizome; parapatao; p e f t o ; p e r i f e r o ,
p e r if e r o m e ; p e r i k i k l o n o , p e r ik i k lo n o m e ; pe ri p la n je m e ; per naoj 2S
pe rpatao ; p e t a o j 2* petjeme/petayomej 2: pi y en o; piyenoerxome;
p i y e n o f e r n o ; p ib a o ; p l e o ; p l i s i a z o ; r i x n o , rixnome; ( s a l e v o ) ;

- 333 -
s e rja n iz o ; (para)serno, (para)sernom e, parasirom e; sikono, sikonome;
sinobevo, sinobevome; ska rfa lon o; (s k iv o ); (sproxno); s triv o j

s u la ts a ro ; ta ksibe vo; tin a z o , tinazom e; (tra va o ); tre xo j 2Z

triyirnao/triyirizoj^Z t s u l a o j 2Z v y a z o , vyeno; v i B i z o , viBizom e;
v u lja z o ; vu ta o j; xa m ilo n o j x i n o , xinome; xono, xonome; (xorevo);

xo ropibao

( e) A ________ (B) se a f t o to sim io

A ________ (B) t o / o n / a t / i n t h i s po i nt

akoluBo, akoluBjeme; a k u b a o j ^ Z anakatevo, anakatevome;

ana pob oy irizo , anapoboyirizome; aneveno, anevazo; anevokatevazo,
anevokateveno; (apoyiono, apoyionome); armenizo; bazo, beno; bi z o,
bizome; b u s u l i z o ; eksfenbonizo, eksfenbonizome; ektoksevo,
ektoksevome; erxome; f e r n o ; f e v y o ; f t a n o ; gremizo, gremizome;
y i r i z o j £ 3; Yl i s t r a o j ^ Z y o n a t i z o ; ipoxoro; ipsono, ipsonome;
kalp azo ; ka t a b io k o , katabiokome; katevazo, kateveno; kaBome, kaBizo;
katra kila o ; k ik lo fo ro ; k ila o j k i l j e m e ; k i n i y a o ; k o l i b a o ; kremao,
kremjeme; ksanaerxome; k s a n a y i r i z o ; ksaplono2f 3Z kunaoj 2 > kunjeme;
k u t r u v a l a o ; kuvalao, kuvaljeme; metafero, metaferome; metakomizo;
obi yo, obiyume; parapatao; p e f t o ; p e r i f e r o , p e r if e r o m e ; p e r i k i k l o n o ,
p e r ik i k lo n o m e ; p e r ip la n je m e ; perpatao; p e t a o j £, petjeme/petayomej 2:
piy eno ; piyenoerxome; p i y e n o f e r n o ; p i ba o; p l e o ; p l i s i a z o j ^ Z
pr os y io no , prosyionome; prospernao; r i x n o , rixnome; sale vo;
s e r j a n i z o ; serno, sernome; sikono, sikonome; sinobevo, sinobevome;
(siom e); s k a r f a l o n o ; s k i v o ; (s pro xno); s t i n o , stinome; s t r i f o y i r i z o ;
s t r i v o j s2'» s u l a t s a r o ; ta k s ib e v o ; tarakunao, tarakunjeme; tara zo,
tarazome; t i n a z o , tinazome; tra dazo, tradazome; ( t r a v a o ) ; tremo;
t r e x o j 2* t r i y i r n a o / t r i y i r i z o j ^ Z tsu la o j^ Z vazo; vyazo, vyeno;
viBizo, viBizome; v u l j a z o j ^ Z v u t a o j 2* vutjeme; xa m il o n o j ^Z x in o ,
xinome; xono, xonome; xore vo; xoropibao

- 334 -
 B unless: even if

LIST
iteration iteration
I : □ □ I : * I : ■
II : ■ II : R|

III:
II : □
III: □ □ □ III: ■ III: ■

A c la s s ific a tio n
tremo x in o gremizo piyeno, 2 piSao sikono
tremble pour pull down, go, take to jump raise/lift
xinome iturl sikonome
p e rife ro e p is tre fo prospernao rise/get up
carry around Siasxizo vu lja z o j 2 return overtake,
p e r if e r o m e traverse sink ' pass by vutao
go around ■f ii riz o p i so dive
xorevo v iB iz o return ta landevo
k ik lo fo ro dance sink oscillate ti n a z o
circulate vidizome erxome talandevome shake up
xor opi Sao f come tinazome
eor o, eorume ,

of
trip riz o hop/jump metakomizo
roam around about move house/ sway ektoksevo
furniture lead 1aunch

Modern
kilje m e metakomizome beno ektoksevome
wal1ow get 1n
s eksfe nbo niz o
p e rik ik lo n o
tfl/strao encircle bazo hurl
slide,, slip p e r ik i k lo n o m e put in eksfenbonizome

Greek
o'bifioj s k a r fa lo n o bi z o rixno
drive climb stick in throw

m eta fer o xono s t e ln o

verbs
m e ta fe ro
transport/ transport/ stick into send
carry over carry over
metaferome m eta fe ro m e VTfeno petao2
go out throw

of
kuvalao kuvalao
carry carry vgazo strivoj 2

motion
ku valjeme kuvaljeme take out turn ’
aneveno aneveno vazo X iriz o j 2
ascend ascend put in/on/at turn ’
anevazo anevazo

along
take up take up kaBizo
sit(TR)
kateveno kateveno
descend descend Y o n a tiz o
katevazo katevazo kneel

the
take down take down
tradazo
ipsono ipsono shake, jolt

process-event
raise raise tradazome
isponome ipsonome Zb
rlis tra o i
xa m il o n o j 2 x a m il o n o j 2 slip
stoop, lo&er stoop, lofter
ksaplono2
siko no sikono lie down
raise/1 ift raise/1 ift
sikonome sikonome fevjfo
rise/get up rise/get up leave

continuum
pernaoj 2 p er n a o j 2 apop ono
pass ’ pass ’ take off
apop'onome
p efto p efto
fall fall petjeme2
jump up
per pat ao per pa ta o
walk walk ( kseki nao )
set out,
trexo trexo start off
run run
salevo
k ila o k ila o stir, move
roll roll siightly
katra k ila o ka tra k ila o kunao
roll down roll down move
petaoj pe ta o 1
fly fly
k o l ibao k o l ib a o
swim swim
2a
serno serno
drag drag
sernome sernome

t a ks ib e v o tak si be vo
travel travel
kin iya o k in iy a o
chase, hunt chase, hunt
akoluBo akoluBo
follow follow
akolu&jeme akolu&jeme

ka ta Sj oko kata Sjo ko

chase, pursue chase, pursue
kataSjokome kataSjokome

sino'bevo sinobevo
accompany accompany
sinoSevome sinoSevome
LIST IV: R e la tiv e a g e n tiv ity o f selected causative Modern Greek
verbs o f motion

VERBS ANIMACY INTENSION CONTROL CAUSE

HUM AN I NAN SI WI NI CC WC NC DC IC NC
SM NSM
V NF

m ta ko m izo 2 + - - - - + + +
move
(furniture)

y lis tra o 2 + - - - - “ + + +
si ide

s triv o 2 + + - - - + + +
t ur n

* The scales are as described in section 3 .3 follow ing G ivoVs ( 1 9 8 4 )

c ire ria , ' I n t e n t i o n ' stands f o r ' v o l i t i o n ' and the scale is:
SI=strong in t e n t , WI=weak in te n t (which I c a ll 'simple i n t e n t ' ) and
NI=non-intent (instead o f Givon's 'n o n -v o lu n ta ry ')
The 'c o n t r o l' scale is :
CC=clear c o n tro l, WC=weak c o n tro l, NC=no control
The 'causation' scale is:
DC=direct cause, I O i n d i r e c t cause and NC=non-cause
'Cause' is used instead o f 'causation'

* HUM=human, AN=animacy, INAN=inanimacy

* SM=self-moving o b je c t, NSM=non-self-moving object

* V=vehicle (or conveyance), NF=natural force or natural phenomenon

- 336 -
VERBS ANIMACY INTENSION CONTROL CAUSE
HUM AN I NAN SI WI NI CC WC NC DC IC NC
SM NSM
V NF

p i 7si a zo 2 + + - - - + +
approach,
bring
close to

epistrefo£ + X - - - + + +
retu rn ,
bring back

ti n a z o + + + - - + + +
shake up

paramerizo2 + X - X - + + X + X
put aside

metafero + + + X - + +
transport

yirizo 2 + + + - X + + +
turn

ipsono + + X - - ~ + ~ +
r a i se

kunao 2 + + + + - + + + - -
move

xa m il o n o j + + X X - + + +
lower

337 -
VERBS ANIMACY INTENSION CONTROL CAUSE
HUM AN I NAN SI WI NI CC WC NC DC IC NC
S'I NSM
V NF

petao 2 + + + + - + + +
throw

r ix n o + + + + X + X + X + - -
throw,
drop

epanafero + - - - - + + + + +
bring back

kila o 2 + X - - - + + X +
ro ll

tradazo + + - X X + + - - +
shake, j o l t

tarazo + X - X X + + +
shake, s t i r

biz o + + + - - + + +
s tic k /
engulf into

viQizo + X + + - + + +
sink

vuljazo2 + X - X - + + +
sink

- 338 -
VERBS ANIMACY INTENSION CONTROL CAUSE
HUM AN I NAN SI WI NI CC WC NC DC IC NC
SM NSM
V NF

k ik lo fo ro 2 + X - - - + + x - + -
c ir c u la te

aposi ro + X - - - + ~ + + + + -
withdraw

anevazo + + + - X + + + + + -
r a is e ,
take up

katevazo + + + - X + + + + + -

take down

sikono + + X + - + + + + + -
lift

p e rife ro + + X - - + + + + + -
carry about

apomakrino + + - + X + + + + + -
take away,
remove

xono + + + - - + + x + x -
s t ic k in to

serno + + + + - + + + + + -
drag

- 339 -
VERBS ANIMACY INTENSION CONTROL CAUSE
HUM AN I NAN SI WI NI CC WC NC DC IC NC
S4 NSM
V NF

s t e l no + + - - - + + + + X

send

fe r no + + + + - + + + + X

bri ng

vazo + + + - X + + + + +
put

vyazo + + + - X + + + + + -
take out

pi yeno 2 + + + - - + + + + +
t a ke to

tr ex o2 + X - - - + + +
make s . o .
run

- 340 -
 LIST V: R e l a t i v e a g e n t i v i t y o f s el e ct e d non- ca us at iv e Modern Greek
verbs o f motion

VERBS ANIMACY INTENSION CONTROL

HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

katabioko + + + - - + - - +
chase, pursue

k in iya o + + + - - + +
chase, hunt,
run a f t e r

ska rfa lon o + + - - - x + +

cl imb

ko lib a o + + - - - X + +
swim

o r mao + + + - X + +
dash, f a l l
v i o l e n t l y on

xorevo + X - - - X + +
dance

* The n ot a ti ons used here are the same as those used in L i s t IV

- 341 -
 VERBS ANIMACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

vaSizo + - - - - -j- + - -
walk, march

porevome + X - - - + +
walk (a long
d i st a n c e )

pezoporo + - - - - + + -
walk, go on
foot

s u la ts a ro + - - - - + +
stro ll, saunter

s e rja n izo + - - - - + +
walk around

b u s u liz o + - - - - + +
crawl (as o f
a baby)

perpatao + + - - - + +

wal k

p e r i p i an j erne + + - - - + X +
wander

\o n a tiz o j + X - - + X + X
kneel

- 342 -
VERBS ANIMACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

ka8ome2 + x + x
sit

ksaplonc>2 + x + x
l i e down

pibao x +
jump, l eap

trexoj + x x +
run

metakomizo
move (house)

salevo
stir, move
s lightly

s triv o j + +
t u rn

\m z o j + +
turn

taksibevo + x
travel

akoluBo
accompany

- 343 -
VERBS AN I MACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

pleo + + + - - + + +
sail

petaoj X + + - - + + + .

fly

p 7Ye/?o + + + - + + + +
go

beno + + + + + + + +
get i n t o ,
enter

vyeno + + + + + + + +
go out

kik lo fo ro j + + + - + + + +
circulate

aneveno + + + + + + + +
ascend

kateveno + + + + + + + +
descend

Sia sx izo + + + + + + + +
traverse

fevyo + + + + + + + + X

leave

- 344 -
VERBS AN I MACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

p lis ia z o j + + + + + + X + + X
approach

y i r i z o ^ ( p is o ) + + + - + + X + + X
return

epi s t r e f o j + + + - + + X + + X
return

ipoxoro + + X + - + + + + X
recede, subside

sikonome + + + + + + +
r i s e , get up

ipsonome X X + + + + + +
r i se,
be e l e v a t e d /
r a i se d

xam ilono + + + + + + + +
stoop, lower

petayomep + + X + + X + +
jump up, be
thrown up

- 345 -
VERBS ANIMACY INTENSION CONTROL
HUM AN I NAN . SI WI NI CC WC NC
SM NSM
V NF

tinazom e + + + X + + X + +
shake/jump up,
be shaken up,
jerk

apomakrinome + + + + X + + + + X
move away, be
removed

e pistrefom e X X X - + + - +
be r etur ned

sernome + + + - + X x + X X -f
crawl

obiyume + + + - - + + - + +
be l e d / d r i v e n

metaferome + + + - - + + - X +
be t r a n s p o r t e d ,
change premises

erxome + + + - + + + + + +
come

xonome + + + X + + +
get s t u c k /
engulfed

- 346
 VERBS ANIMACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

pefto + + + + + x + X +
fall

viBizome + + + - + x + x +
sink

y7 i s t r a o j + + - - + x + X +
s lip, slide

a nevazome + + + - + + +
be r a i s e d /
taken up

ka tevazome + + + - + + - +
be taken down

kuvaljem e + + - - + + +
be c a r r i e d

ku tru valo + + X - - + +
roll down

v u lja z o + + + - + + +
s in k

tremo + + + + + + +
tremble

trandazom e + + + - + + +
be s h a k e n / j o l t e d

- 347 -
VERBS ANIMACY INTENSION CONTROL
HUM AN I NAN SI WI NI CC WC NC
SM NSM
V NF

tarakunjem e + + + - + + +
be shaken

k a tra k ila o X X X - + + +
r o l l down

ta ra zom e - - - + + + +
s tir , be
s t ir r e d

bizome - - - - + + +
g e t s tu c k /
engulfed

k ila o j - - - X + + +
r o ll

- 348 -
LIST VI: Taxonomic sets and natural classes of Modern Greek verbs
of motion

1 k u n j e m e = move [ p a r t ia l m otion]
2 k u n a o = cause to move [ p a r t i a l m otion]
3 p iy e n o j / e r x o m e = go / come
4 perpatao / vabizo = walk
5 fevyo / a n a x o r o = le a v e , d e p art
6 aneveno / a n e r x o m e = ascend [upward m otion]
7 a n e v a z o = r a is e , c a r r y /ta k e up
8 s i k o n o = r a is e , lift
9 kateveno / k a t e r x o m e = descend [downward m otion]
10 k a t e v a z o = b r in g /ta k e down
11 pefto = fa ll
12INTR spevbo / [move r a p id ly ] = hasten
12TR [cause something to move f a s t ]
13INTR [move s lo w ly ]
13TR [cause something to move s lo w ly ]
14 rix n o / p e t a o 2 = throw
15 ferno / p i \ / e n o 2 = b rin g / ta k e to
16 s triv o j / y iriz o j = tu rn [r o t a r y m otion]
17 s triv o 2 / \ i r i z 02 ~ cause to turn
18 [roam around]
19 v i Q i z o m e ~ s in k j^ jp
20 s te ln o = send
21 b e n o = g e t in t o , e n te r
22 v y e n o - go out/come out

23 p i b a o = jump

24 p e r n a o = pass

25 [CL - o b je c t's main p a rt or a l l limbs in co n ta ct w ith ground]

26INTR [r o ll, s lid e ]
26TR [r o ll, s lid e ]
27 (n a v ig a tio n term s)
28 [ta k e away]
29 [b r in g /ta k e back]
30 [go back]

- 349 -
31 [cause something to change lo c a tio n - o b j e c t s main p a rt or a ll
lim bs in c o n ta c t w ith ground]
32 petaoj = fly
33 viQ izo = cause to s in k
34 [ta k e someone o r something here and th e re ]
35 proxoro = advance
35 opisQ oxoro = re tre a t, move backwards
37 v y a z o = ta k e /b r in g out
38 v a z o = put

- 350 -
 CD
rd
+J

rd
Cl

to
fd

CL
fd
■r—
+ ->
lO
J- CD LO
td >
CL o
E
<D
S
cu
fd
to
c:
is CL _Q
-V
CL

to

LO

cr>
co

© CD
> o;
CD o >>
o E
fd 4->
rs* Co «* -C ro
CD o S_ cn 4->
>- •r- U- o
fd 4*4 s*. 4-> i~ o
(o CO fd CO to to

- 351
 o
+->

5
CD
00 + ->
co
CD £_
00 _Q
r3 -r-

00 u

+J
00

lO
o.

o.

DO
+-> +->

+J

LO

03
00

+-> CD r0
4-» +-»
<0 CD <xt
O 00 £_ C l
— =j ja
CD

lio rs

352
 o
N
ra

CD

CD

O O
Q_

-o

CO CD

-CS

+->
V)
o
o
O

o
o
C.
l

353
 +-»

*3-
C\J

tO t_J
Q.
IQ to
_Q 00
CO i— I
CL
o o
o o
nj

cl i— S—
O (/) C
D

t_>
lO o
CO

nJ
V
cn

o
N to to to co
-C to

CO
o OO
o o
“rO
O
to
CO
o
>■

•O

to cr>
*o
cn
cn
co it—

to _Q
£=
CL,

C
D

C CO
O 4-J CO |— Cn
to o -2* cn
cn *»—. CO
CL to to

- 354 -
 ID
c_>
> o

C\J
<-J>
Cl O

CL

Cn
<D 10
co

co

ro
Q.
-O
o.

Ln

C. Cl
Qj a>
cl xj
ro

■O
to

CD

fO

- 355 -
 t. 4-
CL> O
4-> LO
CD
re
5
4—
O

to o CNJ
CD ■— * CNJ
e CD
O CL) CD
CD
LO

ex.
motion]
[upward

Cl

CL
CO X)
CO LO
4 -, -Q Cn J 71
CO CD
0)
anerxome

>
o

c0
ascend

4—1
c-
fO
/

CL CL
aneveno

CL

4->
CL
C 5

LO >1
r—
CL 4^
Q
C5
l CL CO
C
NJ
C. CD
4
-> -Q
CD CO
O

CD -

LO Q.

O
CD

O CL CO
4- =J Q)
O
CD

- 356 -
 M-
-n- cnT
'"H LO
<o CNJ
o Cl
C3
cn

P
<o
CNJ
ITS o
o
SC
ro

Q.
ro
cn o
ro o
up
carry/take
anevazo

Ct
o
raise,

'r—
lO tO

00

4-5

cu

>>

CT)

4-5

<o
to
ro
cu s_

- 357 -
 fd
a)
co
4- td
o QJ cu
CO
ai 4^>
u a>
n3 sz O
CD 4- +-> +->
e c
o =3 o o
c: CO +->
o c
to a) o
fd <— DO
+->
In s jC.
motion]

'CD _c DO
co o •i—
o D i— o
c. O fd +J
cu +->
[downward
katerxome

descend
/
kateveno

- 358 -
 CD
> O
o

o
o
p efto

fa ll
11.

4-
r— CD
CD O
03 C CO c
r»- q ;
CD =s
3S C o
‘r-~ o f— CD o CL
-a z: 5
03 i— i o 3
i- i— CO C; o c
so i— CNJ s- o
03 o O 'K -C CL
s- O s-> =3

*"s
<o CO
o ©
c: cn so CD
<D 5 \— 4, C
CD o 2 : o
<0 - O i—i ^CL
•r*. CO co £Z
CD CNJ o o CO
o •o o CD
S O •r— o O. S-J CD
03 r— CD C=
CO 03 J/

- 359 -
 o
H-J

o
=3
oj S-
ro
QJ o
,H o
o
QJ OJ
O. XJ

Q)
o +o
c: 03
•f— t-
X Cl)
03 1— r°
Ho CD to
'1— O
Cl O
QJ Of
CU
r a p id ly ]

"-I (J
Co -M ■t—
o CO =3
o 03 cr
1— » 4-
CL o
t— " h
X •r— to “O
03 05
[move

o 03
+o CO X CD
QJ -C
x. OJ
QJ ,40
e CD 'O £= Q)
£= 40
/

o +J X =3
c; 03 ex S~ o CO
*r-m V- OJ 03
hasten
spevbo

X 0) t\l E 4 —
Of l“ - o o
40 CD ‘X to QJ
*P*-» CJ *-H
QJ >
Cl. CJ to ■O
X. CD O
QJ 03 o c co
HO CO E
X =3 i— i =S
QJ o «s 03 O
U 4—1
12INTR

x. i- CO
00 HO 03 1— t
*t— . o
to QJ x. C t_J
C-O C QJ =3
O o cx 5-
>
QJ £=
X. QJ
O •-H CT3
> U to =5
•r-» ■ i" o O < M QJ
X. =3 X S-
O H->
> CT QJ sz CM c= OJ
X. <_> •r»« X
o X QJ
c 03 1—
=f HO QJ
S- 'N. O
CX O
© QJ 03
X
QJ £Z
S- =3
HO 5- to _£Z
to c
o 5:
03 o cn
"O o
o o
cr
o ra
+J x
CO
*-H o j
to 4-
o ©
i x. >> N Q_
1o X. k 03 O
Q . Q) Q. r—
> r~- —
1
X 03 03
03 •X Cn
HO
OJ
S

- 360 -
 13INTR [move slo w ly]

- 361 -
 o

o O a>
s. •»->
• 5 ra
CD o o c.
• + 4> o o
1^ xj CL s_
00 £= 00 r-.
fa •i— CD _£Z
CD m L.
CL 5 c o 40 s-
A; o CD o fa CD
CD s- a> <4- c: >
DC JLZ CD ra O
+-> +-> XJ

o
N c
*r~- JZ •s.
cl. 00 o
ra * 1— XJ
— no 1—
CD o 1—
40 E 1—
fa CD 3
-L: XJ CL

00
CD
o o
+-> Cl)
•i—
Q.
JZ
3
+■> c
petaop

CL
>
U o
o X5
throw
/

r0
rixno
14.

o
a X> CJ
r-— . CD
O 03 s-
is- CD Cl
o S 5- 3
s. o CD CL
CD $- JZ
_3 CL
00 4— 5 00

O
r-— .
, o 00
3: CD
o o r~
CD S- O
*f— -3 +4
r->- +-> 00

O
N £_
'h. CD
CL
L_ 4-J
O ro
-V U
Co 00

CD
O X>
ra ■i—
40 00
CD ro
CL
ca +o
00
fa 03
CL U

- 362 -
- 363 -
 £=

>- o3
on
40 00

on
c:
o
o
E "O
>>
s-
03
+J CL
O 3
s-
O

r—1
to TD
o C
3
> o O *ct-
s*. s- O
lO 03 o
o.
r—
t—i +J
i—
03
Cl
00

to

4->

-o

cr

+J

- 364 -
- 365 -
 CD C=
S~ -I-

co as
co ~a

>■ o. cn
as

C_J
O

CJ
LO C_5
O,

00 CD
ro

ra

JQ

CD CD

- 366 -
 •''4
*0
<u
E=
o
N QJ
*f-~ Cn
■4-0 s-
Ci. CD
CO £
X3
=5
<u to

I— ,
O r——I
OJ a) I— i
l+ J > C_J
=J -r- o
^ "O

QJ
JC
CO +^>
<o
o
>*~4 4-> „ —
o £
s- O
ra »l ■■*)
*o C 4-J
Cj ■r— o
4- co -Q

<U C£
e I—
o 2 :
N 1-0
*i— *—1 rs
CD C
■[—
o s-
CO
N QJ
ra *o
*ro
1 c
— 13 c
c> O
CT) 0— LO

QJ
E
.O
lO »— 4
ra QJ i— I
■4-0 > o
ra •r— o
~a

QJ ■O
-Q QJ
o
S- CL)
o QJ S-
1> XJ «:
J3 C a.
lb. 13 *p—
ra O -CZ
c: 4- V)

QJ
B CD
o jC
■N
f—.
+->
Xj O
C: +o
o £
ti­ o
ro +->
■4-0 c + ->
ro •r—
O
CO -Q

- 367 -
 ro
5
o ro
c: CO
>< -a CNJ
o c o
■*-> Q) o
to CO

o
• QJ
o CO 4-
'S QJ
O S-
+J
QJ
ex.
to -o CD
L. £= CO
ro CL) r—
CL CO QJ

O
QJ
to »iClM «o-
-V: a i— l
o c: <_5
HO 3 O
-V rO
QJ r—

c
f “H S-.
Co 3
o
l~~.
40 40
Q) c
QJ S- rO QJ
HO
<■0 c CO QJ
o ■r” ro U
c l ro
to *0 "O ,—
c c Cl
C: QJ QJ QJ
to CO CO S-

QJ
S-
• O
_c 4-
40 QJ
• -Q
CO
'— 1 4-
to • O
0 0
•Q • 40
. 55 CO c
QJ 0
CL S-
to "O 4 -
O sz
Q) c
CD CL CO ■1—
+->
+-> (O
CO 'r“
ro -a
(_> rO
~o S-
ro
O
s- -M
o _Q
CO
1 £3
Q> “a sz
CL c rO
A; CD S-
qj CO 40

- 368 -
 o
c;
CD
■JQ

O \
+J >> +->
C r— - i—
*1— +J 3
CD U
o 5- -i—
Cn U M—
CD U_
o 00 •r—
c: ■c *o
o o C_J>
‘ ex. s~ O «—
s- _ 1 +->
S*. =3 C l.
XJ z : 3

- 369 -
 o

+->

<4-

oo +->

'+J

CD
>- + ->
C

CD

CD

QJ
O.

+->
CD

Ql

c
00

Cl

lO

- 370 -
 qj 4 -
e o
o CL
>- +j 3
03 =3
O cn
3
j Cu CL • i—
<u EE %-
to 3 CL
X : ,rO CO

ro CL
to 3
o
03 CD
3
,X *r—
'r^ s-
CL
to CO

o x
03 Q)
lO >
O
Cl CM
x CL O
o QJ E O
03 EX 3
VO •<o

CO
CM
LO
Cl -M O
O

CL
r—. 03
03 QJ
to r“

O
03 CL
CO 3 QJ
■p* O
EX CL 3
03 E 3
EC 3 O
03 ‘CO _Q

- 371 -
 s-
O Cl)
N >
• r- O

O
O
VO

10

00

VO

CO
Cl
C_>

VO

DO

>,
-Q fO
4->
00 £_
''O CL) CO
10 0 .0 .

ro
OJ

VO

cn

cn
OJ
o

oo
o
VO CL

- 372 -
 CL
QJ
CL>
“O

cn LO
OJ
O
C_>

LO
C_> fO
C_>
LO
o
N
CO
JO lO o
— -s
c? >> o
JO
<TJ
03 X I

ro
<4-
O

CO
ro

Ch O r—
N 2=
ro
S-
'=3 o *=J-
lO o
=3 o
-O

L f)
OJ

ro

CL
o
ro

- 373 -
- 374 -
 XJ CO ,—
’—1 s= to S-
to =3 0 0 CXI
0 O s*. CD xz 0
0 s- fa >
to r~— ca 4. fa
0
a
0
XJ CL CD ca
s- ‘r - 1— Su _c
rd S- •1—
CD ra
0 JX CL to
N U
*r— 03 CO
to -O CO ca
0 0 O
Cl, r—. t_> 0
fa •i—1 4^>
ra *o
c: c
C: <a r—H ra
fa to to 5- CD E Cl
0 ra E
CD CD *r-— O xz
1—- c: O to
Cl
fa
4.
ra ta ta
“c:
O cl to
'-'I CD
<a 4- to JC
<4 . 0 4->
-L;
O E r—— O
*r-
> ra CD 4-5
CD CD
c: CD -Q ra CL
fa to _c -L;
0 4-4 CD \ xz
e XZ O to
0 JX 4-4
L, 4-> -V ta
to •«“ C= r—
O £ O CD CD
CD CD
40 >• ,_ -O fa
ra
5-
ra
CD
rO <a •1— c: -V XJ to
to *—< cn r-— ca •1—
E <0 to 3r
u. 0 >
CD 0 CO
40 1—. c
CL
a
0 to E
•i— 4-, r— p" O
•
to +4
^ __ zs
40 4-4
ro ra *4—^
ra c: to O 0 o
N .o — n -Q
•r— fa
c: r°
CD _
tO CD
•1—
xz
S i +-> ro
_£T !L ca cnj CO
H1 fa to * o
•“" t 4-5 CD
0 © —-
I-'. to s- S- O
C\J 0 xz *d - CD fa O
QJ 4-> 00 4-4 Xj C
r-—
r—
O
C_)
r0 © >r—
CL 4- to
ra •r— CO CD
•r-. fO to _n 4-
10 to O +4 O
N
0 CD ro
t, 4-4 CJ to Cn
CD ra 4-
C CD 4- O 4-
cd E S- L. O
to O =3 fa
-V O to Cl ,— >1 uo
*f— fa o
ra ra 3: o
to to ra

•"I
CO
o CL ro
o s •r— to
CD CD JZ -
r— 4-4 to O
CL
■t-
ra
O IO L-
o CO ta
Cl r— *r— C to 4-
CD 4- HO •r- +4 4-
to 1— 4. 0
*1— O f—
ra
to
ra
*~-t to U-
to >> 0 CD
■■---~ 1— ra 0 -O
*1— 1— _
CD
0 s: >
CD 03 rO CL O
r-— to 10 1— N
CL 1— C_D •1— *r~. zx
O 4-4 ■<— O +-> ta VO u
CL CD (a to to O ra
ta to to 4-5
CL

" 3„75 -
 >>
ro

o ro o
OJ
~a (_>
Q fZ t_ >
Q>
lO to

c
<D s-
(O 3
o r— i— +->
CD 0) 3 o tu
o CL CL tO s-
fC L 0) 0)
to i- S- Q. -
-iC

Cn
away]

< \l
o CL) JQ
N -f-j
[take

•a
■|—. s- ■r—
cn
£, to to cn
QJ CL to £=
E to
to xz
i. +-> to
<v o
28.

to 3
C l. to CL CO

cn
CNJ

+J 3
O 3 to
O s-
*r~ *o
to 3 XZ
O rO 4-J
CL. s- ■i—
(O ■a 3

CL)
Z*Z
o to
c: +J

[C
QJ
ra >
e O >>
O E to
CL.
to
CL) 3
to 3

to >>
o to
1 ^ 3
l*r— to
E
o >1
■V s-
o s-
Q. to
to o

- 376 -
 CD
XJ

c \| CO
o
ro r.

CD

ca cd
S- CD

n—

o
o o
4—5
4-5 CO
-a

CL)
E Cl
O
CO
CD

XJ

•^C XJ CL
r-H
CO
CO
c_>
o

40

- 377 -
 CD
CD
’— s £=
3
r^—
CD
£= CD r — i CL
03 > <o
+o o r>
O
,c l XI N o
•r- x>
re
L. •W sc
<u cl -p—
>3 03
cl
Cl

CD

■o
SC
3 <\1 3 CD
O O •t— C D
o 03 c
pet aoj

%-
-LJ
<0 rts N CL
•r— 1
—
CD *r—i
e +-> CD •O Ol
ro 3 "p~
+0 O
CL X I
•r— ro
*p - .
32.

L. >> .
QJ ro
CL ro

*3 CD
c CD i-
O S- SD
CL CD T0
0J EE c
+o XI 3
03 3 O
A; CO cj—

- 378 -
 CD U-
T3 CD

- 379 -
 u
ro
X) CD
X3

>1
4-> >>
r—
=3
o
<4—
4- <4-
u-
XT rv jr
+ ->
>> +->
■p -
£ £
O -—>
1—
C/1 vo
"O
>1 s-
O S- <0
40 <D £
VO </>
ro > jy ;
TU
U- u LD
0 £ ro O
- c:
O
CD
CO
X l
CO
c_>
QJ £ •p .
ro > ZxZ VO c_>
CJ ro 1----
O O ro 0 ■—
C; CD X I <_> ro ro <0
-V £ £
•*< jy :
proxoro

advance

1— 1—
<13 ro <0
■ i* £ £
35.

vo
X3
s-
rO
?
o
ro I--.
XI C\J
O
C_>

CL
CD
<-0 + J
0 VO
c; \
0 CL) CO
c— > ■0
0 O S-
E ro
\ £
0 JsC
Vo 1— O
*p-» rO ro
Cl. £ XI

- 380 -
 <A

O
"O
c

cn
o
N
03
>- o <4-
c: O
©
'i— . =3
CD o CU t— 1
US +-5 O
r^. •
co 03 o
OS
s-
o

• 4->
3 *QJ C/J t/l
QJ
Cl O C
\ CM o QJ
s- 4 -
O JC CJ
4 -J TD
+->

O *0

O CM
O- o

- 381 -
 QJ
j*:
o
CL

c
T-j
4-J
10
-Q CL CL

CL

t) 4-> t )
3j
cl! 4->
+ ->
•P -P ID
LD
O QJ CM
O C- C_>
B O CJi
00 QJ

CL CL

c
+-s

CD
CL CL

CL

DO

*o
+-> ro
13 QJ
CL
CL

+-> C CL
C

CL

- 382 -
 CHANGE OF POSITION STATIVE

PARTIAL MOTION PARTIAL MOTION MOTION RESULTING IN

NOT RESULTING REGULAR/REPEATED DIFFERENT POSTURE/POINT
IN DIFFERENT OF SUPPORT
POSTURE

kiIjem e talandevome stinome steko, st e kome ( i m e s t i m e n o s )

wallow o s c illa te assume a standing stand
position
k un j e me , p al ome kaBomej ( i m e k a Bi s me no s)
kunaoy vibrate kaBomeg s i t , be seated
move, s t i r s i t down
t re/no ksaplonoj
sale vo tremble ksaplonog ( ime k s a p l o m e n o s )
s tir 1 ie down l i e , be lying down
klibonizom e
tr ad az o me ro ll and pitch kr emj eme kremome ( ime kremasmenos)
shake, je rk hang hang, be hanging
s 7o/ne
anaoevome , shake \ernc>2 yernoj ( i m e y er men os)
a nakat evome, lean on lean, be leaning
( ana)tarasome 1ik niz ome
s t i r , churn, swing, rock s i k on ome e o r um e , m e t e o r i z o m e ( ime m e t e o r o
shake up stand up be dangling, hanging above
kim atizo
anapibao wave anakaQome epipleo
jump up s i t up f lo a t , be afloat

yonatizoj ime a kub i sm e no s

kneel be leaning on, be on

s k i vo
bend
 BIBLIOGRAPHY

Aitchison, J. (1983), The A rtic u la te Mammal: an In tro d u c tio n to

P s y c h o lin g u is tic s , London: Hutchinson ( 1 s t edn 1976)

A l e xi o u, M. (1982), ' D i g l o s s i a in Greece' , in W. Haas ( e d . ) , S t a n d a r d

Languages, Spoken and W r i t t e n , Manchester: Manchester U n i v e r s i t y
Press'

Allerton, D . J. (1982), V a le n c y and the E n g lis h Verb, New York and

London: Academic Press

Anderberg, M.R. (1973), C lu s te r A n a lysis fo r A p p lic a tio n s , New York:

Academic Press

Anderson, O.M. ( 1 9 7 1 ) , The Grammar of Case: Towards a L o c a lis tic

T h e o r y , Cambridge: Cambridge U n i v e r s i t y Press

Anderson, J.M. ( 1 9 7 3 ) , An E s s a y C o n c e r n i n g A s p e c t , The Hague: Mouton

Apresjian, U u. D. , Me l 'c u k, I.A. and ^olkovskij, A.K. ( 1 9 69 ) ,

'Semantics and l e xi cogr aphy : towards a new type of unilingual
dictionary', in K i e f e r (1969)

Babiniotis, G, ( 1 9 7 2 ) , To Ri ma t i s E llin ik is , Athens: V ivliothiki

Sof ia s N. Sar ipol ou

Babiniotis, G. and Kontos, P. (1967), S y g c h r o n i k i Grammatiki tis

K o i n e s N ea s E l l i n i k i s , Athens

Bach, E. and Harms, R. (eds.) (1968), U n ive rsa ls in L in g u is tic

Theory, New York: H o l t , Ri nehar t and Winston

- 385 -
Baumgartner, K. (1967), 'Die Structur des B e d e ut un gs f el d es ' , in H.
Moser (ed.), S atz und Wort im H eutigen D eutsch, Mannheim:
I n s t i t u t f u r Deutsche Sprache

Beard, R. ( 1 9 8 1 ) , The I n d o - E u r o p e a n L e x i c o n , Amsterdam: North Holland

Bendix, E.H. (1966), C o m p o n e n tia l A n a l y s i s o f G e n e ra l V o c a b u la r y : The

Semantic S tru ctu re of a Set of Verbs in E n glish , H ind i a nd

J a p a n e s e , Bloomington: I ndi ana U n i v e r s i t y

Bennett, D.C. (1972), 'Some observati ons concerning the locative -

directional d i s t i n c t i o n ' , S e m i o t i c s 5, p p . 58-94

Bennett, D.C. (1975), S p a tia l a nd Temporal Uses of E n glish

P re p o s itio n s : an Essay in S tra tific a tio n a l S em antics, London:
Longman
\

Berlin, B. (1971), ' S p e cu l a t i on s on the growth of ethnobotanical

nomencl atur e' , L an g u a g e - B e h a v i o r R e s e a r c h L a b o r a t o r y P a p e r 39,
Berkeley: U n i v e r s i t y o f C a l i f o r n i a Press

Berlin, B. , Breedlove, D. and Raven, P. (1968), ' Cov er t c at egor ies

and f o l k t ax onomi es' , A m e r i c a n A n t h r o p o l o g i s t 70, p p . 290-9

Berlin, B. , Breedlove, D. and Raven, P. (1973), 'General principles

of classification and nomenclature in folk biology', American

A n th ro po log ist 75, p p . 214-42

Berlin, B. and Kay, P. (1969), B asic C o lo r Terms: T h e ir U n iv e rs a lity

a nd E v o l u t i o n , Berkel ey: U n i v e r s i t y o f C a l i f o r n i a Press

Bi er wi sch, M. (1971), 'On classifying semantic features', in

Steinberg and Jakobovits (1971)

- 386 -
Bi nni ck, R. (1971), ' Br i ng and come', in L in g u is tic In q u iry 2,
p p . 260-65

BogusTawski, A. ( 1 9 7 0 ) , 'On semantic p r i m i t i v e s and meani ngf ul ne ss' ,

in A . J . Greimas, R. Jakobson, e t a l.(ed s .), Sign, Language,
C u l t u r e , The Hague: Mouton

Bostanjoglou, Th. (1962), A n t i l e x i k o n i Onom astikon tis N e o e llin ik is ,

Athens: Domi

Brown, R. ( 1 9 7 3 ) , A F i r s t L a n g u a g e , London: A l l e n and Unwin

Brown, G. and Yul e, G. (1983), D iscourse A n alysis, Cambridge:

Cambridge U n i v e r s i t y Press

Carnap, R. (1952), 'Meaning postulates', P h ilo s o p h ic a l S tu d ie s 3,

p p . 65-73

Chase, W.G. and C l a r k , H.H. ( 1 9 7 1 ) , 'Semantics in t he per cepti on of

v e r t i c a l i t y ' , B r i t i s h J o u r n a l o f P s y c h o l o g y 62, p p . 311-26

Chomsky, N. (1965), Aspects of the Theory of Syntax, Cambridge,

Mas s. : MIT Press

Chomsky, N. (1968), L an g u a g e a nd M ind, New York: Harcourt Brace

(Enlarged edn 1972)

Chomsky, N. (1972), S tu d ie s on Semantics in G enerative Gr ammar , The

Hague: Mouton

Chomsky, N. ( 1 9 7 6 ) , R e f l e c t i o n s on L a n g u a g e , London: Temple Smith

Chomsky, N. ( 1 9 8 1 ) , Lectures on G o v e r n m e n t a n d B i n d i n g , Dordrecht and

Cinnaminson: For is

- 387 -
Clark, E.V. (1974), 'Normal s t a t es and evaluative viewpoints',
L a n g u a g e 50, p p . 316-32

Clark, E.V. (1978), ' Di s co ve r in g what words can d o ' , in P a p e r s f r o m

the P a rase ssio n on the L exicon, Chicago: Chicago Linguistic
S oci ety

Clark, E.V. and Garni ca, O.K. (1974), 'Is he coming or going? On the
acquisition of deictic verbs', Journal of Verbal Learning a nd

V e r b a l B e h a v i o r 13, p p . 559-72

Closs Traugott, E. (1978), 'On the expression of s pat io-te mpor al

relations in l a n gu ag e ', in Greenberg, Ferguson and Moravscik
(1978)

Cole, P. (ed.) ( 1 9 8 1 ) , R a d i c a l P r a g m a t i c s , New York: Academic Press

Cole, P. and Morgan, J.L. (eds.) (1975), Syntax and Semantics 3:

S p ee c h A c t s , New York: Academic Press

Coleman, L, and Kay, P. (1981), ' P r o t ot y pe semantics: the English

word 7 / e ' , L a n g u a g e 57, p p . 26-44

Comrie, B. (1976), Aspect: an In tro d u c tio n to the Study of Verbal

Aspect and R elated P roblem s, Cambridge: Cambridge University
Press

Comrie, B. (1981), Language U nive rsa ls and L in g u is tic T ypo log y,

Oxford: Basil Blackwell

Coser iu, E. (1971), Sprache: S trukturen und Funktionen, Tubingen:

Gunter Narr

Coseriu, E. (1973), Probleme der S tru c tu re lle n S em antik, Tubingen:

Gunter Narr

- 388 -
Coseriu, E. and G eckel er , H. (1981), Trends in S tru ctu ra l Sem antics,

Tubingen: Gunter Narr

C r e s s w e l l , M. (1978), 'Prepositions and poi nts o f v i e w ' , L in g u is tic s

a nd P h i l o s o p h y 2, p p . 1-41

Cruse, D.A. (1973), 'Some thoughts on agentivity', Journal of

L in g u is tic s 9, p p . 11-23

Cruse, D.A. (1977), 'The pragmatics o f l e x i c a l s p e c ific ity ', Journal

o f L in g u is tic s 13, p p . 153-64

Dahlgren, K. (1978), 'The natur e of linguistic stereotypes', in

Papers from the P a rasession on the Lexicon, Chicago: Chicago
L i n g u i s t i c Soci ety

Darden, B. , Bailey, C. and Davidson, A. (eds.) (1968), Papers from

the Fourth M eeting of the Chicago L in g u is tic S o c ie ty , Chicago:
U n i v e r s i t y o f Chicago, Department o f L i n g u i s t i c s

Davidson, D. and Harman, G. (eds.) (1972), Sem antics of N atural

L a n g u a g e , Dordrecht: Reidel

Denny, P.O. (1978), ' L o ca t i n g t he u ni v e r s a l s of spatial d eixis', in

Papers from the P arasession on the L exicon , Chicago: Chicago
L i n g u i s t i c Soci et y

Di k, S.C. ( 1 9 7 8 ) , F u n c t i o n a l G r am ma r , Amsterdam: North Holland

Di mi tr a ko s , D. (1969), Neon O rthographikon F rm in e u tiko n L exikon ,

Athens: Giovanis

Dixon, R.M.W. (1971), 'A method of semantic description', in

Steinberg and Jakobovi ts (1971)

- 389 -
Dowty, D.R. (1972a), 'On the syntax and semantics of the atomic
p r e d i c a t e CAUSE', in P a p e r s from the E ighth R egional M eeting of

the Chicago L in g u is tic S o c ie ty , Chicago: Chicago Linguistic

Soci ety

Dowty, D.R. (1 9 72 b) , S tud ie s in the Logic of Verb Aspect a nd T i me

Reference in E n g lish , Ph.D. D i s s e r t a t i o n , U n i v e r s i t y o f Texas a t

Austin
D o w t y , D. R. ( 1 9 7 9 ) , W o r d Meaning and Montague Grammar ,
Dordrecht: Reidel
Dowty, D.R. (1982), 'Grammatical r e l a t i o n s and Montague grammar', in
P. Jacobson and G.K. Pullum (eds.) (1982), The Nature of
S y n t a c t i c R e p r e s e n t a t i o n , Dordrecht: Reidel

Dummett, M. ( 1 9 7 3 ) , F r e g e : P h i l o s o p h y o f L a n g u a g e , London: Duckworth

Eikmeyer, H. d. and Ri e se r , H. (eds.) (1981), Words, W orlds and

C o n t e x t s , Berlin: De Gruyter

Ferguson, C.A. (1959), ' D i g l o s s i a ' , Wo r d 15, pp. 325-40

Fillenbaum, S. and Rapoport, A. (1971), S t r u c t u r e s in the S u b je c tiv e

L e x i c o n , New York and London: Academic Press

Fillm ore, C. J. (1966), 'Deictic c at e gor i es in the semantics of

"come"', F o u n d a t i o n s o f L a n g u a g e 2, p p. 219-27

Fillm ore, C. J. (1968a), 'Lexical entries for verbs', Foundations of

L a n g u a g e 4, p p . 373-93

F i l l m o r e , C. J. (1 9 68 b) , 'The case f o r c a s e ' , i n Bach and Harms (1968)

Fillm ore, C. J. (1969), 'Types of lexical information' in Kiefer

( 1 9 69 ) ; also in Steinber g and Jakobovits (1971)

- 390 -
Fillm ore, C. J. ( 1 9 71 a ) , ' E nt a i l m e n t r u l e s in a semantic t h e o r y ' , in
Rosenberg and T r a v i s (1971)

Fillm ore, C.J. (1 971 b) , 'Some problems for case grammar' in R.J.
O'Br ien (ed.) (1971), Georgetown U n iv e rs ity Monograph S e rie s on

Languages and L in g u is tic s 24, Washington D. C . : Georgetown

U n i v e r s i t y Press

Fillm ore, C. J. (1971c), 'Verbs o f j udgi ng: an e xe r c i s e in semantic

description', in C. J. F i l l m o r e and D.T. Langendoen, ( e d s . ) (1971),
S tu d ie s in L in g u is tic Sem antics, New York: Holt, Ri nehar t and
Winston

Fillm ore, C. J. (1975), 'An alternative to checklist theories of

meaning', in Proceedings of the F irs t Annual M eeting of the

Berkeley L in g u is tic s S o c ie ty , Berkeley: Berkeley Linguistics

Soci ety

Fillm ore, C. J. (1978), 'On the o r ga n i z at i o n o f semantic i nf or ma t io n

in t he l e x i c o n ' , in P a p e r s f r o m the Parasession on the Lexicon ,
Chicago: Chicago L i n g u i s t i c Soci ety

Fodor, J .A . (1970), 'Three reasons for not deriving "k ill" from
"cause to d i e " ' , L in g u is tic In q u iry 1, p p . 429-38

Fodor, J .A . (1972), 'Troubl es about a c t i o n s ' in Davidson and Harman

(1972)

Fodor, J . A . , Bever, T.G. and G a r r e t t , M.F. (1974), The P s y c h o l o g y o f

Language: an In tro d u c tio n to P s y c h o lin g u is tic s a nd G enerative

Gr ammar , New York: McGraw-Hill

Fodor, J . A . , Fodor, J . D . and G a r r e t t , M.F. (1975), 'The psychological

unreality of semantic representations', L in g u is tic In q u iry 6,
p p . 515-31

- 391 -
Fodor, J.A ., Garrett, M.F., Walker, E. C. T. and Parkes, C.H. (1980),
' Aga ins t d e f i n i t i o n s ' , C og nitio n 8, p p . 263-367

Fodor, J . D . ( 1 9 7 7 ) , S e m a n t i c s , New York: Crowell

Frajzyngier, Z. (1982), 'In d e fin ite agent, passive and impersonal

passi ve: a f u n c t i o n a l study', L i n g u a 58, p p . 267-90

Geach, P. and Black, M. (eds.) (1966), T r a n s l a t i o n s from the

P h ilo s o p h ic a l W ritin g s o f G o t t l o b F r e g e , Oxford: Blackwell

Geckel er, H. (1971), S tru c tu re lle Semantik und W o rtfe ld th e o rie ,

Miinchen: W. Fink

Givon, T. ( 1 9 7 9 ) , On U n d e r s t a n d i n g G r am ma r , New York: Academic Press

Givon, T. (1984), Syntax: a F u n ctio n a l-T y p o lo g ic a l In tro d u c tio n

v o l.l, Amsterdam/ Phi ladel phi a: John Benjamins

Greenberg, J.H. (ed.) (1963), U nive rsa ls of Language, Cambridge,

Mass. : MIT Press

Greenberg, J. H. Ferguson, C.H. and Moravscik, E.A. (eds.) ( 1 9 78 ) ,

U n ive rsa ls of Human Language 3 : Word S t r u c t u r e , S t a nf o rd , C a l.:
Stanford U n i v e r s i t y Press

Greimas, A.J. (1966), Semantique S tru c tu ra le : Recherche de meth ode,

P a r i s : Larousse

G ri c e, H.P. (1975), ' Logi c and conversation', in Cole and Morgan

(1975)

G ri c e, H.P. (1981), ' Presupposi ti on and conver sat ional im plicature',

in Cole (1981)

- 392 -
Gruber, J.S. (1965), S tu d ie s in Lexical R e la tio n s , MIT Doctoral
Dissertation, Cambridge, Mass. (Reproduced 1970, I ndi ana
University Linguistics Circle)

Gruber, J.S. (1976), L e xica l S tru ctu re s in S yntax a nd Sem antics,

Amsterdam: North Holland

Gunderson, K. (ed.) (1975), Language, M ind and Knowledge,

Mi nne apol is , Minnesota: U n i v e r s i t y o f Minnesota Press

Gunderson, K. and Maxwell, G. (eds.) ( 1 9 75 ) , M innesota S tu d ie s in

P h ilo so ph y of Science 6, Mi nne apol is , Minnesota: University of

Minnesota Press

Halle, M., Bresnan, J. and M i l l e r , G.A. (eds.) (1978), L in g u is tic

T h e o r y a nd P s y c h o l o g i c a l R e a l i t y , Cambridge, Mass.: MIT Press

Hayes, J . R . (ed.) (1970), C og nitio n and t h e D e v e lop m e nt of Language,

New York: Wiley

Hopper, P. and Thompson, S. (1980), 'T ra n s itiv ity in grammar and

discourse', L a n g u a g e 56, p p . 251-99

Hopper, P. and Thompson, S, (eds.) (1982), Syntax and Sem antics 15:

S tud ie s in T r a n s i t i v i t y , New York and London: Academic Press

Householder, F., Kaza zi s, K. and Koutsoudas, K. (1964), Reference

Gr ammar of L ite ra ry D h im o tik i, Bloomington: I ndi ana University

(Publication 31 of the Indi ana University Research Center in
Anthropology, Folklo re, and L i n g u i s t i c s )

Hudson, R.A. (1980), S o c io lin g u is tic s , Cambridge: Cambridge

U n i v e r s i t y Press

- 393 -
Hunn, E. ( 1 9 8 2 ) , 'The u tilita ria n factor in f o l k biological
c l a s s i f i c a t i o n ' , A m e r i c a n A n t h r o p o l o g i s t 84, p p . 830-47

Ikegami, Y. ( 1 9 6 9 ) , The S e m o l o g i c a l S t r u c t u r e o f t h e E n g l i s h V e r b s o f
M otio n , Tokyo: Sanseido

Oackendoff, R.S. (1972), Sem antic In te rp re ta tio n in G enerative

G r am ma r , Cambridge, Mass.: MIT Press

Jackendoff, R.S. (1975), 'Morphological and semantic r e g u l a r i t i e s in

the l e x i c o n ' , L a n g u a g e 51, p p . 639-71

Jackendoff, R.S. (1976), 'Toward an e xp l ana tor y semantic

representation', L in g u is tic In q u iry 7, p p. 89-150

Jackendoff, R.S. (1978), 'Grammar as evidence for conceptual

structure', in H a l l e e t a l . (1978)

Jackendoff, R.S. (1981), 'On K a t z ' s autonomous s em an ti c s' , L an gu ag e

57, pp. 425-35

Jacobson, P. and Pullum, G.K. (eds.) (1982), The M a t u r e o f S yn ta ctic

R e p r e s e n t a t i o n s , Dordrecht: Reidel

Johnson, S.C. (1967), 'H ierarchical clustering schemes',

P s y c h o m e t r i k a 3 2, p p . 241-54

Johnson-Laird, P.N. (1982), 'Formal semantics and the psychology of

meaning', in S. Peters and E. Saarinen (eds.), Processes,
b e l i e f s , a n d q u e s t i o n s , Dordrecht: Reidel
J o h n s o n - L a i r d , P.N, ( 1 9 8 3 ) , M e n t a l M o d e l s ,C a m b r i d g e :
Cambridge Universi ty Press
J ohnson-Laird, P.N. and Wason, P.C. (eds.) (1977), T h i n k i n g : R eadings
i n C o g n i t i v e S c i e n c e , Cambridge: Cambridge U n i v e r s i t y Press

- 394 -
Katz, J.J. (1966), The P h i l o s o p h y of Language, New York: Harper and
Row

Katz, J . J . ( 1 9 7 2 ) , S e m a n t i c T h e o r y , New York: Harper and Row

Katz, J.J. (1975), ' Logi c and language: an examination of r ecent

c r i t i c i s m s o f i n t e n s i o n a l i s m ' , in Gunderson and Maxwell (1975)

Katz, J.J. (1977), 'The real s ta tus of semantic representations',

L in g u is tic I n q u i r y 8, p p . 559-84

Katz, J . J . (1980), 'Chomsky on meaning', L an g u a g e 56, p p . 1-41

Katz, J.J. and Fodor, J. A. (1963), 'The structure of a semantic

theory', L a n gu a g e 39, p p . 170-210

Katz, J.J. and Nagel, R .I. (1974), 'Meaning p os t ul at es and semantic

t h e o r y ' , F o u n d a t i o n s o f L a n g u a g e 11, p p . 311-40

Kay, P. and McDaniel, C.K. (1978), 'The l i n g u i s t i c sig nificance of

the meanings o f basic c o l o r t e r m s ' , L a n g u a g e 54, p p . 610-46

Kazazis, K. ( 1 9 8 2 ) , 'P artial linguistic autobiography of a

s ch i zo g l o s s i c l i n g u i s t ' , G lo ssolog ia 1, p p . 109-17

Keenan, F. L. ( 1 9 7 2 ) , 'On semantically based grammar', L in g u is tic

I n q u i r y 3, p p . 413-61

Kenny, A. (1963), A c tio n f Emotion a nd M ill, New York: Humanities

Press

Kempson, R.M. ( 1 9 7 5 ) , P re s u p p o s itio n a nd the D e lim ita tio n of

Sem antics, Cambridge: Cambridge U n i v e r s i t y Press

- 395 -
Kempson, R.M. (1977), Sem antic Theory, Cambridge: Cambridge
U n i v e r s i t y Press

Kiefer, F. (ed.) (1969), S t u d i e s in Syntax a nd S e m a n t i c s , Dordrecht:

Reidel

Kiefer, F. and Muni tz , M.K. (eds.) (1970), Language3 B e lie f and

M e t a p h y s i c s , New York: S t at e U n i v e r s i t y o f New York Press

Ki m ba l l , J.P. (ed.) (1975), Syntax and Sem antics 4, New York and
London: Academic Press

K ri pk e, S.A. (1972), 'Naming and n e c e s s i t y ' , in Davidson and Harman

(1972)

Labov, W. (1978), ' De n o ta t i o n a l structure', in P a p e r s from the

P arase ssio n on t h e L e x i c o n , Chicago: Chicago L i n g u i s t i c Society

L a ko f f , G. ( 1 9 7 1 ) , 'On generative s em an ti c s' , in Steinber g and

Jakobovi ts (1971)

Lakoff, G. (1972a), ' L i n g u i s t i c s and na tur al lo g ic', in Davidson and

Harman (1972)

Lakoff, G. (1 9 72 b) , 'Hedges: a study in meaning c r i t e r i a and the

logic of fu zz y c o n ce p ts ' , Journal of P h ilo s o p h ic a l Logic 2,
p p . 183-228

Lakoff, G. ( 1 9 7 7 ) , 'Lin gu istic gestalts', in P a p e r s from the

T h irte e n th R egional M eeting of the Chicago L in g u is tic S o c ie ty ,

Chicago: Chicago L i n g u i s t i c S oci ety

La ko f f , G. and Ross, J .R . (1972), 'A note on anaphoric i sl ands and

causatives', L in g u is tic In q u iry 3, p p . 121-5

- 396 -
Langacker, R.W. (1975), ' F u nc t io na l stratigraphy', in P a p e rs f r o m th e
P a r a s e s s i o n on F u n c t i o n a l i s m , Chicago: Chicago L i n g u i s t i c S oci ety

Langacker, R.W. (1982), 'Space grammar, analysability, and the

English passive', in L a n g u a g e 58, p p . 22-80

Lascaratou, C. and Philippaki-Warburton, I. (1981), 'The use of

passive c o n st r uc t io n s in Modern G r ee k' , in M andatoforos 17,
p p . 53-64

Leech, G.N. (1969), Towards a Semantic D e scrip tio n of E n glish ,

London: Longman

Leech, G.N. ( 1 9 7 1 ) , M e a n i n g a n d t h e E n g l i s h V e r b , London: Longman

Leech, G.N. ( 1 9 7 4 ) , S e m a n t i c s , Harmondsworth: Penguin

Lehr er , A. (1970), ' S t a t i c and dynamic elements in s e ma n ti c s' , P a p e r s

in L in g u is tic s 3, p p . 349-73

Lehrer, A. (1974), Sem antic F i e l d s and L e x i c a l S tru ctu re , Amsterdam:

North Holland

Lehrer, A. and Le hr er , K. (eds.) ( 1 9 7 0 ) , T h e o r y o f M e a n i n g , Englewood

C liffs, N J: P r e n t i c e - H a l l

Levelt, W.M. (1970), 'A s c a l i n g approach t o the study o f syntactic

relations' in G.B. Fl ores D' Ar ca is and W.J.M. Levelt (eds.)
(1970), Advances i n P s y c h o l i n g u i s t i c s , Amsterdam: North Holland

Levinson, S.C. (1983), P ragm atics, Cambridge: Cambridge University

Press

- 397 -
Long, B.S. ( 1 9 7 5 ) , The D e v e l o p m e n t o f S e m a n t i c F e a t u r e s : How C h i l d r e n
Learn Verbs of M otio n , unpublished Ph.D. Dissertation, Cornell
University

Lounsbury, F.G. (1954), 'The structural analysis of kinshi p

s e ma nt ic s' , in H. Lunt (ed.), in P roceedings of the N inth
In te rn a tio n a l C o n g r e s s o f L i n g u i s t s , The Hague: Mouton

Lyons, J. ( 1 9 6 3 ) , S t r u c t u r a l S e m a n t i c s , Oxford: Blackwell

Lyons, J. (1968), I n t r o d u c t i o n to T h e o re tica l L i n g u i s t i c s , Cambridge:

Cambridge U n i v e r s i t y Press

Lyons, J. ( 1 9 7 7 ) , S e m a n t i c s , 2 v o l s . , Cambridge: Cambridge U n i v e r s i t y

Press

Lyons, J. (1981), Language, Meaning and C ontext, Bungay-Suffolk:

Fontana

Mackridge, P. ( 1 9 8 5 ) , Modern Greek, Oxford: Oxford U n i v e r s i t y Press

McCawley, J . D . (1968a), 'The r o l e o f semantics in a grammar', in Bach

and Harms (1968)

McCawley, J .D . ( 1 96 8 b ) , 'Lexical insertion in a t ra n s f o r m at i o n a l

grammar w i t h ou t deep s t r u c t u r e ' , in Darden e t a l . (1968)

McCawley, J .D. ( 1 9 71 a ) , 'Meaning and the d e s c r i p t i o n o f l anguage s' ,

in Rosenberg and T r a v i s (1971)

McCawley, J.D. ( 1 971 b) , 'P re le x ic a l syntax', in R. J. O' Bri en (e d.)

(1971), Georgetown U n iv e rs ity Monograph S e rie s on Languages a nd

L in g u is tic s 24, Washington D. C. : Georgetown U n i v e r s i t y Press

- 398 -
Mel ton, A.W. and M a r t i n , E, (eds.) (1972), Coding Processes i n Human

M em or y , Washington D . C . : Winston

Mereu, L. (1983), 'Is t h er e such a th i ng as a l e x i c o n in models of

linguistic comprehension?', paper read at the Second

In te rn a tio n a l C olloquium on the In te rd is c ip lin a ry Study of the

Semantics of N atural Language, M e a n i n g a nd t h e L exicon , Kleve:

Nijmegen I n s t i t u t e o f Semantics (ms.)

Me rvis, C.B. and Rosch, E. (1981), 'Categorization of na tur al

o b j e c t s ' , A n n u a l R e v i e w o f P s y c h o l o g y 32, p p . 89-115

M iller, G.A. (1969), 'A psychological method t o investigate verbal

c on ce p ts ' , J o u r n a l o f M a t h e m a t i c a l P s y c h o l o g y 6 , p p . 169-91

M iller, G.A. ( 1 9 7 1 ) , ' E m p i r i c a l methods in the study of s emanti cs' ,

in Steinber g and Jakobovi ts (1971)

M iller, G.A. ( 1 9 7 2 ) , ' E ng l i s h verbs of motion: a case study in

semantics and l e x i c a l memory', in Melton and M a r ti n (1972)

M iller, G.A. and Johnson-Laird, P.N. (1976), L a n g u a g e a nd P e r c e p t i o n ,

Cambridge, Mass.: Harvard U n i v e r s i t y Press

Montague, R. (1970),'English as a formal l a n gu ag e ', in Vincentini

(ed.) (1970), L in gu ag g i N e lla S ocieta e N e lla T ecnica, Mi lan:
E di z i o n i di Comunita

Mourel atos, A. (1978), 'Ev ent s, processes and s t a t e s ' , L in g u is tic s

and P h i l o s o p h y 2, p p . 415-34

Moore, T. and C a r l i n g , C. (1982), On U n d e r s t a n d i n g Language: Towards

P o s t - C h o m s k y a n L i n g u i s t i c s , London: Macmillan

- 399 -
Morley, G,D. (1983), 'A g e n tiv ity : a componential f un c t i o n a l
approach', L i n g u a 60, p p . 177-82

Nida, E.A. (1975), Com ponential A n alysis o f Meaning: an I n t r o d u c t i o n

t o S e m a n t i c S t r u c t u r e s , The Hague: Mouton

Osherson, D.N. and Smith, E.E. (1981), 'On the adequacy o f prototype
t he or y as a t heory o f c once pt s' , C o g n i t i o n 9, p p . 3 5- 58

Paul, P. (1982), 'Homonyms, semantic divergence and v a l e n c y ' , Lingua

58, p p . 291-307

Perlmutter, D. (e d.) (1983), S tudies in R e la tio n a l Grammar 1,

Chicago: U n i v e r s i t y o f Chicago Press

Peterson, P.L. (1982), 'Anaphoric r e f er enc e to facts, propositions,

and e v e n t s ' , L in g u is tic s a n d P h i l o s o p h y 5, p p . 235-76

Peterson, T.H. (1983), 'Semantic s t r u c t u r e ' , Journal of L in g u is tic s

19, p p . 79-114

Pulman, S.G. ( 1 9 8 3 ) , Wo r d M e a n i n g a nd B e l i e f , London: Croom Helm

Putnam, H. (1970), 'Is semantics possible?', in Kiefer and Munitz

(1970)

Putnam, H. (1975), 'The meaning o f meaning', in Gunderson and Maxwell

(1975)

Quine, W.V. ( 1 9 6 0 ) , M o r d a n d O b j e c t , Cambridge, Mass.: MIT Press

Quine, W.V. (1971), 'The inscrutability of r e fe re n c e ', in Steinberg

and Jakobovits (1971)

- 400 -
Rhodes, R. (1983), ' L e x i c a l taxonomi es', paper read at the S e c on d

In te rn a tio n a l C olloquium on the In te rd is c ip lin a ry Study of the

Semantics of N atural Language, Meaning a nd the Lexicon, Kleve:

Nijmegen I n s t i t u t e o f Semantics (ms.)

Rips, L .J., Shoben, E. J. and Smith, E.E. (1973), 'Semantic d i st an ce

and the verification of semantic relations', Journal of Verbal
L e a r n i n g a nd V e r b a l B e h a v i o r 12, pp. 1-20

Rosch, E. (1973), 'Natural Categories', C og nitive Psychology 4,

p p . 328-50

Rosch, E. (1977a), 'Linguistic re la tiv ity ', in Johnson-Laird and

Wason (1977) (First published in A. S ilverstein (ed.) (1974)
Human Com m unication: T h e o re tica l P e rsp e ctive s, Hillsdale, N J:
Laurence Erlbaum A s s o c i a t e s . )

Rosch, E. ( 1 977 b) , ' C l a s s i f i c a t i o n o f r e a l - w o r l d o b j e c t s : o r i g i n s and

representations in c o g n i t i o n ' , in Johnson-Laird and Wason (1977)
( F i r s t published in S. Ehrlich and E. T ul v ing (eds.) (1976), La

memoire s e m a n t i q u e , P a ri s : B u l l e t i n de P s y c h o l o g i e . )

Rosch, E. (1977c), 'Human c a t e g o r i s a t i o n ' , in N. Warren (ed.)

(1977), S tu d ie s in C ro ss-C u ltu ra l Psychology, London: Academic
Press

Rosch, E. (1978), 'Principles of categorization', in Rosch and Lloyd

(1978)

Rosch, E. and Lloyd, B. (eds.) (1978), C og nitio n and C a t e g o r i z a t i o n ,

H i l l s d a l e , N J: Lawrence Erlbaum Associates

Rosch, E. and M er vi s, C.B. (1975), ' Fa mi ly resemblances: s tudi es in

the internal structure of categories', C o g n itive Psychology 7,
p p . 573-605

- 401 -
Rosch, E., Me rvis, C.B., Gray, W.D., Johnson, D.M. and Boyes-Braem,
P. (1976), ' Ba s ic obj ect s in natural categories', C og nitive
P s y c h o l o g y 8, pp.382-439

Rosenberg, J. and Travis, C. (eds.) (1971), Readings in the

P h i l o s o p h y o f L a n g u a g e , Englewood C l i f f s , N J: P r e n t i c e - H a l l

Ross, J. R. (1972), 'A ct', in Davidson and Harman (1972)

Sampson, G. (1975), The For m of Language, London: Weidenfeld and

Nicol son

Sampson, G. (1979), 'The in d ivisib ility of words ', Journal of

L in g u is tic s 15, p p . 39-47

Sampson, G. ( 1 9 8 0 ) , M a k i n g S e n s e , Oxford: Oxford U n i v e r s i t y Press

S a va si r , I. and Gee, J. (1982), 'The f un c t i o n a l e q u i va l en t s o f the

middle voice in child l an guage ', in Proceedings of the E ighth
Annual M eeting of the B erkeley L in g u is tic s S o c ie ty , Berkeley:
Berkeley L i n g u i s t i c s Soci ety

Schweidenweiler, F. (1942), 'Die W ortfeldtheorie', Z e its c h rift fu r

deutches A lte r tu m 79, p p . 249-72

S eiler, H. (1968), 'Probleme der V e r b - S u b k a t e g o r i s i e r u n g ' , L i n g u a 20,

p p .337-67

Shibatani, M. (1973), 'Semantics of Japanese causativization',

F o u n d a t i o n s o f L a n g u a g e 9, p p . 327-73

Shibatani, M. (ed.) ( 1 9 76 a ) , Syntax a nd Sem antics 5: Japanese

G e n e r a t i v e G r a m m a r , New York and London: Academic Press

- 402 -
Shibatani, M. (ed.) ( 1 976 b) , Syntax and S e m a n tic s 6: The Grammar of

C a u s a t i v e C o n s t r u c t i o n s , New York and London: Academic Press

Shibatani, M. (1985), 'Passi ves and related c on st r uc t io n s: a

protot ype a n a l y s i s ' , L a n g u a g e 61, p p. 821-48

Sondheimer, N.K. (1978), 'A semantic a na ly s is o f r e f e re n c e t o s p a t i a l

properties', L in g u is tic s a n d P h i l o s o p h y 2, p p . 235-80

Stalnaker, R.C. (1968), 'A t he or y of conditionals', in N. Rescher

(e d.) (1968), S tu d ie s in L og ica l Theory, Oxford: Oxford
U n i v e r s i t y Press

Stalnaker, R.C. (1972), 'Pragmatics', in Davidson and Harman (1972)

S te in be rg, D. and J ak obov it s, L. (eds.) (1971), Sem antics: an

In te rd is c ip lin a ry Reader in P h ilo s p h y , L in g u is tic s and

P s y c h o l o g y , Cambridge: Cambridge U n i v e r s i t y Press

Talmy, L. (1972), Sem antic S t r u c t u r e s in E n glish a nd A t s u g e w i , Ph.D.

D i s s e r t a t i o n , U n i v e r s i t y o f C a l i f o r n i a at Berkeley

Talmy, L. (1975), 'Semantics and syntax o f m ot i o n ' , in Kimball (1975)

Talmy, L. (1976), 'Semantic c aus at ive t y p e s ' , in S hi bat ani (1976b)

Theophanopoulou-Kontou, D. (1983), 'Patient vs non patient

orientation o f the a ct ion and the voice d i s t i n c t i o n in MG', in
G lo ssolog ia 2-3 ( 1 9 8 3 - 1 9 8 4 ) , p p . 75-90

Thomason, R. (1974), Formal P h ilo s o p h y, New Haven: Yale University

Press

Thomason, R.H. and Stalnaker, R.C. (1973), 'A semantic theory of

a dve rbs ', L in g u is tic In q u iry 4, p p . 195-220

- 403 -
Trier, J. (1931), Der Deutshe W ortschatz im S in nb ezirk des
V e r s t a n d e s , Hei del be rg: C. Wi nter

Tz ar tza nos, A. (1946) N e o e llin ik i S yn ta xis (tis K o inis D im o tik is )

(2nd edn), Athens: Organismos Ekdoseos Scholikon V i v l i o n

Vendler , Z. (1957), 'Verbs and times', P h ilo s o p h ic a l Review 66,

p p . 143-60

Verschueren, 0. (1981), 'Problems o f l e x i c a l s ema nt ic s' , Lingua 53,

pp. 317-51

Warburton, I. (1975), 'The passive in English and G r ee k ', F o u n d a t i o n s

o f L a n g u a g e 13, p p . 563-78

Wei nrei ch, U. (1963), 'On the semantic structure of l anguage', in

Greenberg (1963)

Wei nrei ch, U. (1966), 'Explorations in semantic theory', in T.A.

Sebeok (ed.), Current Trends in L in g u is tic s v o l . 3, The Hague:
Mouton

Werner, H. and Kaplan, B. (1963), Symbol Form ation; an O rganism ic-

D evelopmental Approach to L an g u a g e and t h e E x p r e s s i o n o f T h o u g h t ,

New York: Wiley

Whorf, B. L. (1956), 'Language, Thought and R e a l i t y ' , i n d. B. Carroll

( e d . ) , S e l e c t e d W r i t i n g s o f B e n j a m i n Lee W h o r f , Cambridge, Mass.:
MIT Press

Wi er zbi ck a, A. ( 1 9 7 2 ) , S e m a n t i c P r i m i t i v e s , F r a n k f u r t : Athenaum

Williams, E. ( 1 9 8 1 ) , 'On the notions " l e x i c a l l y r e l a t e d " and "head o f

a w or d" ', L in g u is tic In q u iry 12, p p . 245-74

- 404 -
Wittgenstein, L. (1953), P h ilo s o p h ic a l In v e s tig a tio n s , New York:
Macmillan

Wotjak, G. ( 1 9 7 1 ) , U n t e r s u c h u n g e n z u r S t r u k t u r d e r B e d e u t u n g , Berlin:
Akademie (2nd edn 1977)

Zwicky, A*M, and Sadock, J.M. (1975), Ambiguity tests and how to
fail them', in Kimball (1975)

- 405 -