Idea Transcript
L 'O'! 'J11' l
Ruth
1.
In traduction There have
A.
Tel
Aviv
a
number
been
I.....
•
Form and Function: Middles, and Impersonals in Modern
Passives,
•
IS LS I 'n-q
Hebrew*
Berman
University
of
recent
studies
on
imperson
al passives in relation to personals and to other types of impersonal constructions (for instance, Comrie 1977, Keenan
to
appear,
1978,
Kirsner
Perlmutter
&
1976,
Postal
Langacker
1977).
[1]
per is to structions
investigate three in Modern Hebtew,
related and to
tinctions
between
terms
them
in
&
Munro
The
1975,
purpose
Olshtain
of
this
types of agentless characterize the
(1)
of
structural
pa con dis
factors
conditioning their formation and (ii) of pragmatic factors affecting when they are used and/or how they are interpret ed. It seems that detailed examination tions in a given language might suggest correlates between the existence of als and middle voice, on the one lack of resorting to the use of language does have a productive passive-formation.
of such construc certain typological
well-developed imperson hand, and a concommitant passives even where a structural mechanism for
The three Hebrew constructions in question are illus trated in (1) and (2) below: The first two constructions passives of in
and
middles
-
share
constituents, differing the fact that the passive
specified.
The
third,
the
same
surface
SV{X)
ordering
formally in verb morphology and alone may have a overt agent
impersonal
construction
takes
the
form V(X), and differs from both the passive and the middle in
must (1)
the be
morphological in
Type
the
I:
of
the verb,
AGENTLESS
PERSONAL
putax
be
angliya
the-idea
was-developed
in
England
II:
MIDDLE
hara'eyon
the-idea Type
VOICE
hitpateax
developed
III:
3RD
PERSON
(REFLEXIVE)
be
angliya
in
England
PLURAL
INTRANSITIVES
(ACTIVE)
IMPERSONALS
pitxu
et
hara'eyon
be
angliya
developed(PL)
OM
the-idea
in
England
Type
ney two
I:
which moreover
PASSIVES
hara'eyon Type
(2 )
pattern
plural.
[2]
PASSIVE
cvatim teams
ye'urgenu will-be-organized
bekarov soon
BBrke!y Linguistics Society 5. 1979
"--------------
•
2
Type
II:
ney
..
MIDDLE
cvatim
yit1argenu
two teams beka rov
will-get-(themselves)-organized
soon
Type
III:
IMPERSONAL
ye I arg enu
(they)-will-organize The
three
a1 ternates
"ney
cvatim
bekarov
two
teams
soon
in
(I)
and
cally equivalent, in the sense postional content and having the of
the
will all
three
expressions to
In
sives of
the
superficially
extent
the
ble
which
Type
is
I
event
action
in
semanti
same pro And none
they ,
can
the
Yet,
above
or
question:
"agentless".
function
what
do
as
imply
traditional
here,
we
is
not
of
being
the
topic
interest,
of
the
is,
will
they
try
to
respect
or are
show,
to
the
agency_
role
"the
perpetrated
that
differently with or
manifested:
from
(2)
specifies overtly who
perpetrate
such
are
of expressing the same truth value.
of
the
identity (and)
is
agentless
of as
sentence"
the far
(R.
pas
performer as
\
possi
Lakoff
1971:
159), and "the personal passives ••• foreground the NP which is their subject" (Keenan Ms. 61). In Hebrew, as in English (though
by
no
means
cally can specify 'at-the hands-of = laka
'by
(means
---- In the
all
languages),
agency - e.g., by scientists'
of)
Type
such
the-department'.
II
middle
expressions
[3]
voice
intransitives,
specification is much less vague: the verb morphology - both hitpateax
it is in (1)
in
hitpa'
with
(2)
ject
Type or
being
the of
II
the
be -
feasible,
for
a.
that
it
is
also
which
hara'eyon
attached some
kind
the
its
hitpateax
'teams [4]
is
of
by
"reflexive"
i tsel f
idea own
hacvatiml teams
The
dative
with
the
expression
is
"by" I
quite
instance: acmo
both
the
no
from
surface
in
developed' Type
me
the
Thus,
Ob
contrast
developed
'the
pattern
will-get-organized',
hitpateax
the
verb
(nonoccurrent)
agent.
the-idea its
agent
expressed within and y!!'argenu
I the-idea
hara'eyon
of
In
action
but
'the
b.
intransitive
y!"!1 argenu
can
passives
(3)
the
implication
examples,
cvatim
phrase
in
clear
typi
in I of (1) bidey mad'anirn and of (2) hamax
teams
developed
of
itself,
on
its
own,
accord'
yit'argenu will-organize(INTRANS)
will
cases,
(go
the
subject NP marker le-
and)
object
get
themselves organized.'
pronominal
- feyon on the
lahemi to-them
must
agree
with
'idea' or cvatim 'teams'. pronoun object on (3b)
I
I
'-
3
'.
represents a productive type on intransitive verbs - e.g.,
'they
of intensifying hem haIxu lahem
reflexive bli lomar
form mila
went to-them without sayingwmeans something
like 'they went off, they upped and went themselves' or with a middle voice verb, hi hitpatxa yafe metod 'she developed to-her very nicely' has the sense of 'she's gone
and
developea--(herself)
very nicely'.
This reflexivity of
sense is to be expected for Type II middles, for the action is predicated of a coreferential Object and Agent or Ex periencer together. This can explain why Hebrew uses the same verb pattern for middle-voice and for the small set of lexicalized reflexive-verbs referring to bodily activities
(e.g.,
hitraxec
hitgaleax
And
in
'wash
'shave
fact
equivalence
oneself', hitlabe 'dress oneself',
oneself')
as
discussed
many
languages
between
reflexives
tives, reflecting the close these categories (see Barber
show
an
and
in
Berman
to
appear.
overt morphological
middle-voice
intransi
semantic correspondence 1975, Faltz 1977, and
between Garcia
1975). Type III expressions take the form of a Main Verb in 3rd Person Masculine Plural with its associated complements and
modifiers.
hence
are
struction 1973),
a
They kind
(the
and
"impersonal"
have
of
latter
they
no
surface
"verb-first"
term
is
due
constitute
constructions
SUbject
or
to
the
all,
and
"missing-persons"
at
con
Hakulinen
&
prototypical
in Modern Hebrew.
Kartunnen instance
[5]
of
Examples
include: (4)
a.
pitxu et hara'eyon be angliya developed(PL) OM the-idea in England 'They developed the idea in England'
b.
ye'argenu
od
cvatim
will-organize(PL) more 'More teams are going to
c. bonim
am
are-building
'Theylre
there
building
A new roadls d.
Neuter,
new
new
roadl
built
there'
ba
bayit
hahu
kvar
¥anim
not(have)
lived
in
house
that
already
years
house
Elsewhere gender
road
a
shortly'
garim
Nobody's
the
being
xada
10
IThat
why
kVi
bekarov
teams soon be organized
main
hasn't
lived
(Berman
verb
of
(Masculine in
in
that
lived house
in progress),
such
here
fact);
been
why
in/ for
we try
constructions
being
years'
must
functionally
it must
be
Plural
-
to be
demonstrate Masculine
equivalent in
keeping
t the "gener ic" Sense of such proposi tions; and unspecified agent or agents must be construed as
this connection,
see
also
Kirsner
1976).
Below
we
why Human
in
to wi th
the (in
present
4
evidence ly
to
demonstrate
motivated
verb
-
which
in
Hebrew
surface subject in structions cannot to
their
Subject (6)
overt like
such a
on
as
despite the
in
other
with
some
zero subject
'people'
three
of
the
languages
overt,
as anaim 'people'
nominal;
these
form
semantical "impersonal"
agrees
with
the
number, gender, and person such con be analyzed as transformationally related
counterparts
show,
that
constraints
"
that
nor
is
cannot
is, it
it
though
or hem 'they'.
II impersonal"
As
(5)
be -cQreferential
neither
to
pronominalizes
pronominalizable
by
a
a
pronoun
and
any noun like
I they' 6 (5 )
anal(im
ke
a,
b. c.
d,
•
hem
mit'almim
that,
they're
ignoring
toe- root-of
the-problem
ks;,
me'ikar
omr 1m zot, haba'aya
me'ikar
haba I aya
0
k¥:e
anaKim
• ke
The
zot,
say
haba'aya
0
me'ikar
shows
omIlm
people me'ikar when
mit'alrnim
omrim
zot,
0
mit'almim
omrim
zot,
hem
mit'nlmim
haba'aya
contrast
that
0
between wellformed
even
such
a
(a)
"generic"
and
deviant
kind of noun
as
$f)
anam
'people' requires an overt pronominal anaphor, and cannot be taken to cover the same scope of sUbjecthood as the zero subject in (b). As a corollary, hem 'they' must be anaphor
ic
to
anaim 'people'
such
as
politika'im
and
cannot
be
taken
plural verb - as below ill of
is him
im
a.
if
ke
the
c.
most
if as
likely
yedidav subject
(a)
'his of
a
friends') subjectless
example
the people who those who tell
in
situation),
thus:
meraxlim
alav,
hu
mitragez
gossip
about-him,
he
gets-mad
0
mesapr im
10
tell (PL)
him
im
0
meraxlim
hem
mesaprim
rneraxlim 0
mesapr im
Thus expressions like those in tences of (5) and (6) - with
=
alav,
'when
(6)
\
do the talking him about it
people
0
ke
or
implied
Similarly, the
im
ke •
(d).
the
ana¥im
when
b.
as
only well formed are not the same
(pragmatically (6 )
in
(or to some more clearly specified N
'politicians'
he's
told'
hu
mi tragez
hu
mi tragez
10 alav, 10
(4) above and in both main clause
the and
(b) sen embedded
zero subjects - are truly "impersonals" and they do not have any kind of deep subject hem associated with them; for hem 'they' is typically a personal, that is referring, pronoun,
I
,
5
'.
whereas sense
here
of
The
event
sonals,
by
underlying
has
subject
Munro
no
by
(1975:
the
passives
or
in
English
usage
in
fn.
a
5)
by
the
(as -
use
noted
for
the
property of
wide
including
III
in
sive
main
referential
of
in
unspecified",
verb
agent
this
range
both
latter
requiring
study
as
is an
proceeding
as
a
in
the
Type
as
an
on
gen
the
kind
restricted
sub
Biblical
which
constituent
has as
the sur
than impera in which one
constructions
more
imper
the
impersonal
comments
language
of
III of
given universe of typically translat
overt
impersonals
well
too.
a
of
outside
sentences other true of Hebrew,
subjectless
active
impersonals, Before
of
"they"
Gesenius'
face grammatical subject in all tives. This is certainly not
finds
is
794).
of human beings entailed by a (6] Such sentences are quite
ject
peculiar
and
specified
then,
eral class discourse.
ed
"the
Langacker
(see
fn.
5),
labelled
Type
set
ps
of
(7J to
a
consideration
of
the
functional
impact of the three constructions - Type I agentless pas sives, Type II middle voice, and Type III active impersonals we consider structural factors constraining the formation of one or other of these constructions. 2. Structural Constraints mation Although Hebrew has a mechanism
highly
Middle-Voice
productive
- as both
passives
and
Type
II
middles
A rough chart system of verb appear, fn. 2)
and
are
a
morphological pro
on formal grounds. tions of the binyan part from Berman to
passives,
For
ductive means for constructing middle-voice expressions indicated schematically in (7) below - it appears that I
constructing
and
relatively
Type
for
on Passive
severely
constrained
indicating the main func morphology (adapted in is given below:
(7) Main Functions of the Binyan Ver-patterns --
1.
---
kal
-
Basic,
nonderived
tive patax
1-:1
2.
nif'a1
-
-- --i) ii)
both
transi
& intransitive, e.g., caxak 'open', gamar 'finish'------
'laugh',
Intransitive verb
Intransitive,
nirdam
'fall
pattern,
pattern:
middle
(8]
reflex
asleep',
'melt'
Ei' - Banonderived
verb
niftax
of hifil:
'be e.g.,
ni'aemain',
-- --
pattern,
typically
transitive: e.g., gibe 'finalize', piteax 'develop',
3b.
verb
Passive reflex of kal: e.g., opened', nigmar 'be-linished' names
3.
---
'al - Passive
irgen
reflex
'organize'
of verbs
in El'el:
e.g.,
gu
ba 'was finalized', putax 'was developed'
6
4.
hitpa' el
- Intransitive verb pattern: [8) Intransitive, middle reflex of Eilel:
i)
hitgabe
'get
idevelop'
ii)
Reflexive:
hif'il
--
--
Reciprocal:
-
Transitive verb
i )
Causative
'make
of
hirdim
(behind) ,
hof'al
Passive
This
chart
in
hilbi
reflex
of
kal:
shows
that
sleep,
reflex
of
verbs
the
basis
hicxik
(someone) I
in
nif'al:
e.g.,
hTSTir- 'leave
-- --
(intransitive):
e.g.,
hexvir
ripe'
hif'il:
in
e.g.,
(by X)', "hi:irdam I be put
sleep' on
'correspond
e.g.,
'dress
verbs
to
hulba 'be dressed to
oneself',
(one another)'
'turn pale', hivil 'become
-
--
'wash
pattern:
'put
Inchoative
e.g.,
hitpateax
hitkatev
verbs
laugh',
Transitive
iii)
--
e.g.,
(with)', hitnaek 'kiss
ii)
5b.
hitraxec
hitlabe 'dress oneself'
iii) 5.
e.g.,
finalized',
of
transitive
verbs
in
(I) kal, (3) £l'el and (5) hif'il - passives can be formed in (2) nif'al and in (3b) 'al and (5b) huf' respeccive lYi while middle-voice intransitives can be formed in (2) nif'al and in (4) hitpa'el. Below we consider different kinds of constraints - syntactic, morphological, and lexical - on such formations. (a) Only Direct Object Passives: Hebrew allows only DO (Pa tient
to
or
Experiencer)
appear,
promised
Ms.
25-7.
help),
passives,
The
lack
in
of
Oblique passives
the
10
sense
passives
of
I
\
Keenan,
(John
was
(He'll be laughed at)
as well as Instrumental, Locative, etc. passives in He brew is due to a more general syntactic constraint in the language which disallows dangling or orphan tions (as discussed in Berman 1978, 124-238).
forms
like
gossiped'
yicaxek are
constrained without
because
any
examples
they
associated
typically used (b)
'will-be-laughed'
morphologically
instead
interpreand than are
in a due
cally brew
participial in general
sives
like
analogous ing woman
would NP.
of
or but
ruxal
entail
Type
expressions
III
'was
syntactically prepositions
impersonals
like
the
are
English
above.
Present Tense Statals: rather (terms
feasible
preposi That is,
those
to at
hence
Present
used
kinetic or to Hasegawa
tense
passives
in
(8)
to
dynamically passive 1968), reflecting the
nature of present-tense (Berman 1978: 142-59).
English The around :00
tend
be
in a statal or adjectival
below are
verbs That
ambiguous
in
sense basi
in is,
He pas
a
way
room is swept (i) the clean every morning - dynamic passive
I I
7
-
compared
statal
or
ambiguous (i)
(8 )
with
(ii)
so
perfective
constructions
pula!
pattern
a. hagiyot ??
be'adom
are-indicated
in-red
ha'orex
by
the
editor
.mesumano t
livdok
otan
so{it's)
easy
to-check
the
them
kol
hamo'amadim
more
kol
tal for (e)
the-candidates banose
to-deal
with-the-subject
detail
in
are
in
forms
-
terpart and
Berman
in
two
examples
the
same
159-68),
interpreted
indicate
type
these
as
and
Passive
Forms:
gerunds -
have
be
as
set
out
in
the
or
sta
preferred
Nonfinite
no
chart
in
present-tense
verb
passive
purely passive patterns,
hof' al
(as
discussed
adjectival
III impersonals will of the verb. (10J
on Non-finite
the
(a)
of
1978:
typically
infinitives
(5b)
the
examples
- and again, Type the dynamic sense
Constraints
are-qualified
la f asok
further
passives
musmaxim
are-qualified the-candidates a1 ydey va'ada by (a) committee hamo'amadim rnusmaxim
question-marks as
pattern
all
well
now
similarly
in-red,
kal
huf'al
down
be'adom,
are-indicated
az
all ??
As
carpet
Examples of
Hebrew are:
ydey
mistakes
b.
in
the
[9]
a1
b. hag i yo t
a.
can
mesumanot
mistakes
(ii)
we
passive.
coun
(3b) in
'
(7).
'l'hus
the equivalents-of English the needs to be fixed or he
sighed
need an
to
be
with
given
impersonal
relief
an
type
on
his
active
article' being
formulation -
corresponding
to
very
'(they)
completed
often need
to
of fix
the pipe'. As a result of the lack of passive infini tive forms in the two exclusively passive verb-patterns - (3b) ' and (5b) hof'al by contrast with (2) nif'al in
the
ing
an
chart
in
infinitive
again
take
over.
sions
like
the
(7)
above
are Thus
affair
the
must
helped will take the Type (9)
a.
-
used,
when modal
the
active
expressions
tak
impersonal
will
Hebrew equivalents
be
ended
or
IIr-impersonal
xayavim
lesayem
et
must(PL)
to-end
OM
haparaa the-affair
he
of
expres
can't
form: --- -
be
8
b.
(d)
10
yexolim
lesayea
10
not
can(PL)
to-help
him
Lexical Gaps in passive Patterns: There is quite a large group of verbs with respect to which the paradigmatic is sup Active Eil / Passive I / Middle hitpa'el pletive in actual usage; in such cases the passive form, though
morphologically
placed
by
true
passive
take
an
(1 )
( i )
in
agent a.
regular,
the middle-voice some
cases
phrase)
PASSIVE
but
*
MIDDLE
PASSIVE
*
not
in
in
(10)
below
it
re
as
a
can
all. kuba1 was-received
yadeynu) us)
hamixtav
hi tkabe1
the-letter
was-received
(by a.
(e.g.,
being
functions
the-letter
(a1 (i i)
avoided,
which
hamixtav (a1 (by
b.
is
form -
yadeynu) us)
hu
bUka
(a1
he
was-asked
(by
ydey
hamenahe1) the-boss)
lehitnacel to-apologize
b.
MIDDLE
hu
hitbake
(a1
he
was-asked
(by
ydey
hamenahe1) the-boss)
lehitnacel to-apologize For
reasons
which
are
as
yet
not
clear
to
suppletive paradigms admit of no passive all with certain verbs, as evidenced by the (b) examples below are wellformed agent
(11)
phrase
(i)
is
a. PASSIVE * hai' ur (a1
such
suyam
ydey
(by
b. MIDDLE
PASSIVE
*
was-ended hamore) the-teacher)
hai'ur
histayem
the-lesson
was-ended
*
hamore
al
ydey
by
a.
some
possible.
the-lesson
(ii)
us,
construction at the fact that just in case no
the-teacher
hamexonit
teku1ka1
the-car
wi11-be-broken
(a1
hace'irim)
im
titen
the-kids)
if
you-give
ydey
(by ota
it
(to
them)
9
b.
MIDDLE
hamexonit
titkalkel
the-car
wil1-break-down
*
(a1
ydey
(by
hace'irim)
im
titen
the-k ids)
if
you-give
ota
it These
are
of Type sives and
cases
where
(to
Type
I passives middles as in
II
them) middles
are
used
instead
functioning as dual-purpose pas (10b) or as middles alone as in
(11 b) • (e)
Gaps
in
patterns and
the System charted
regular
form
of
stive (12 )
the
set
(i i)
(4)
ACTIVE
Active/Middle
pattern
hitpa'el. pi'el:
alternations
in
Ellel
intran
its
the
dan
sider
et
ha I inyan
Dan
arranged
OM
the-matter
pileI:
hamare
siyem
et
hai' ur
the-teacher
ended
OM
the-lesso
histader the-matter arranged-itself/ got-settled/worked-out
hitpa'el:
less
hal!ii' ur
histayem
the-lesson
ended
productive
alternation
pattern
taking
ACTIVE
hif'il:
is
but qui te manifested by
the
intransitive
haxom
hemis
et
the-heat
melted
OM
widespread the
transi
nif'al
--
--
as
haxem I?
the-butter
MIDDLE
nif' al:
haxem1a
namesa
( baxom)
the-butter
mel ted
ACTIVE
h if' il :
himix
( in-the-heat)
hamore
the-teacher
continued
hai I ur
the-lesson
MIDDLE nif'al: hai'ur verbs
alternate
in
this
et
OM
v
nimsax
the-lesson
Numerous
and
Thus:
its middle-reflex, thus:
( i i )
verb
productive
ACTIVE
hif'il
(i)
(3)
of
highly
ha I inyan
somewhat
(13)
system
one
hitpa'el:
Active/Middle tive
The
includes
MIDDLE
MIDDLE
A
of
Middles: (7)
transitive
reflex ( i)
of in
way
continued
in
Hebrew,
and
the
fact that each such pair has an associated passive form (in .E.'!'al for the verbs like those in (12), in huf' for verbs like those in (13») is evidence for the claim that Hebrew manifests a tripartite system of voice - ac tive, passive, and middle .. [11) However, this system breaks down with respect to verbs in pattern (1) on
10
chart cludes
(7) - the basic both transitive
sitive
(one-place
As
a
the dual
predicate)
result of
distribution
of
or unmarked kal pattern which in (two-place predicate) and intran
both
verbs
verbs.
the
in
dual
the
transitive/intransitve
kal
pattern
function of the nif'aattern
passive
reflex
of
verbs
in
kal
and
the
(1)
(2)
-
along
middle
both
with
reflex
the
of
verbs in hit'il as shown in T7T - no special form exists as the middle counterpart of kal, and niflal may func tion as both the intransitive-middle and--Passive sense in
such
(14)
cases:
( i)
PASSIVE
nibar
hakad
nif'al:
was-broken
the-vase
(al
nif'al:
MIDDLE
( ii)
PASSIVE
toe-cat) nisbar,
the-vase
broke,
nif'al:
Ie
da'avoni
to
my-regret
hadelet
niftexa
the-door
wasopened
(al MIDDLE
haxatul)
ydey
(by hakad
ydey
;
haso'er) the-doorman)
(by hadelet the-door
nif'a!:
1
niftexa opened
pit'om suddenly Again,
Type
where
III
no
agent
is
impersonals
specified,
in
the
active
speakers
kal
can
use
in
such
pattern
cases, too, to yield, for instance, gamru et kol ha'oxel '(they) have-finished OM all the-food' in the sense of
'the food's all finished'
or
have-broken OM
the-vase'
The
constraints
charted (15 )
structural as
=
et
hakad
'(they)
noted
in
this
section
can
be
follows:
Constraint
I
Passives
a) b)
Only
c)
No
DO
d)
Lexical Gaps in Passives Gaps in Set of Middles
Pres.
Passives Tns. Passive
Statal
e)
avru
'the vase has broken'.
II
Middles
Passives
Impersonals
+
-
Nonfinite
III
+
-
+
-
+
(+ )
+
-
(+ )
where minuses indicate that a construction is formally una vailable, pluses indicate the construction used instead, and parentheses indicate that a construction may but need not be used
to
bridge
the
gap.
11
The the
picture
most
which
severely
emerges
is
constrained
that of
Type
the
I
passives
three
are
constructions;
Type II middles are somewhat constrained - as shown by (e), but they may take over the role of passives as in (d); while active
plural
impersonals
constraints
at
formed
intransitive
with
account
also
all
-
of
being
provides
Type
the
and
a
III
only
seem
ones
transitive
formal
to
manifest
which
can
verbs
explanation
of
freely
alike.
This
the
intuitive
feeling of Hebrew speakers that the passive is atypical of their language, somehow "not really despite
the
relatively
which exists for its by the findings of a
productive
formation. pilot study
Tel Aviv University [12] of year-old Israeli children, who the
passive
This the
avoidance was particularly seven subjects, and was to
respect tival
to
our
children
when
produce
point
used
are
children
used
for
those
showed
impersonals
avoid
the
the or
III)
only
mechanism
or
the
called
viewed
NP
the
as
Moreover,
passive
up of
for.
logical
adjec
the
stra
when
required
object
occurred
here:
younger
made
middle-voice
of
backed Ariel
the younger of throughout with
are
points
somewhat Hebrew",
l and one-half avoided using
clearly
[13]
the
command
while
quite
which
the
actives
(Type
passive;
which
with
to
evident in manifested
above).
in
better
used
was
avoiding
consistent
simple
it
8
consistently
passives
(b)
sentences
initially who
even
present-tense
(see
tegies
form
morphological
This intuition is conducted by Mira seven
no
be
(Type
II);
passive
constructions
fronting
strategies
two
oldest
to
children,
who made most frequent use of passive formation (including passives with agent-phrases), made relatively rare use of such methods
of
of
avoidance.
This ties in with another relevant structural property Hebrew, as a language for which the passive is relatively
not .'tAu,,,"I ,q o