See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/320167543

Arabic ‘labile verbs’ in form III: Theoretical and typological perspectives

Chapter · September 2017

DOI: 10.1075/tsl.120.10let

CITATIONS READS

2 135

1 author:

Alexander Letuchiy
National Research University Higher School of Economics
48 PUBLICATIONS 224 CITATIONS

SEE PROFILE

All content following this page was uploaded by Alexander Letuchiy on 19 September 2022.

The user has requested enhancement of the downloaded file.

 y
pan
Arabic ‘labile verbs’ in form III

Com
Lability or something else?

ing
Alexander Letuchiy
National Research University, Higher School of Economics, Moscow

lish
Pub
he article discusses labile (also known as ambitransitive) verbs in literary
Arabic. I show that, though Arabic does not have a rich system of labile verbs,
some existing cases of lability (labile verbs of form III) are particularly inter-
ins
esting from the typological point of view. heir unusual property is that the
opposition between semantically causative vs. non-causative uses (and number
of syntactic arguments) do not always correlate with syntactic (in)transitivity.
am

Verbs of form III have both possible types of non-standard uses: transitive non-
causative and intransitive causative uses, as well as the two standard uses (transi-
enj

tive causative and intransitive non-causative uses). I link this non-standard type
of lability to some general features of voice and transitivity in Arabic. At the end
nB

of the article, some cases of lability outside form III are discussed.
Joh

1. Introduction
-

he tendency of languages to choose either valency increase or valency decrease

to be grammaticalized has been discussed by theoreticians and typologists for a
long time. For instance, Nichols et al. (2004) propose a classiication of the world’s
ofs

languages based on their way of marking transitivity and valency change. hey also
show that these language types are areally and genetically motivated. While in some
pro

linguistic areas, such as Caucasus and South America, valency increase (illustrated
by causativization in this article) tends to be grammatically marked, some other
regions (e.g. Europe) show an opposite tendency: transitive verbs tend to be basic,
ted

while intransitive lexemes are derived by means of valency decreasing operations

(most characteristically, anticausative).
rec

According to Nichols and her co-authors, as well as according to studies by

Haspelmath (1993), Letuchiy (2006, 2010) and others, lability (also known as am-
cor

bitransitivity – ability of verbs to be transitive or intransitive with no additional

morphological marking) is also genetically and areally motivated. For instance, in
Un

doi 10.1075/tsl.120.10let
© 2017 John Benjamins Publishing Company
258 Alexander Letuchiy

y
Europe and Africa lability is much more widespread than in Native American lan-

pan
guages. However, as follows from Letuchiy (2010), in all areas labile verbs are found.
Asian languages do not behave uniformly in this respect: alongside languages with

Com
large classes of labile verbs, there are those which have almost no labile verbs.
Arabic (both the literary language and spoken dialects) belongs to the latter
class. Labile verbs are found here only sporadically. Although they are sometimes
mentioned in linguistic studies, such as Fassi Fehri (2012), more oten authors

ing
of descriptions do not analyze labile verbs in detail. For instance, Benmamoun
et al. (2010) analyze syntactic behavior of transitive verbs, including their ability

lish
to form passives and tests for subjecthood, but do not mention labile verbs. Saad
(1982: 34–54) treats some verbs as transitive and intransitive at the same time if they

Pub
have a transitive use which can be passivized and an intransitive non-passivizable
one. However, his examples of transitive and intransitive verbs are cases of argu-
ment omission rather than cases of lability. He notices that verbs like ’akala ‘eat’
ins
behave as transitive when used with an object but manifest intransitive properties
if the object is omitted. his type of phenomenon is labeled ‘A-lability’ by Dixon
am

(1980), Kibrik et al. (2001) and others, though its typological distribution difers
signiicantly from that of lability proper or P-lability (seeLetuchiy 2006). A-lability
enj

is much more widespread and sometimes characterizes the whole verbal system
of a language.
he present paper is focused on a subclass of Arabic labile verbs, namely, verbs
nB

of form III denoting symmetrical states like ‘be equal’, ‘be close’, ‘be similar’ and
so on (most of them also have a dynamic reading, such as ‘get closer’). hese verbs
Joh

behave in a speciic way regarding their syntactic transitivity. hey are very close
to the notion of lability employed by Nichols et al. (2004), Kulikov (1999), Letuchiy
(2013), Kulikov & Lavidas (2014) in that the same form has both a causative use
-

(e.g. ‘make equal, compare’) and a non-causative one (‘be equal’). At the same time,
these verbs difer from the traditional concept of lability by their syntactic proper-
ofs

ties: their causative uses are not always transitive, and non-causative uses are not
necessarily intransitive. Some verbs of the group have a transitive non-causative
pro

use and an intransitive causative use. For instance, in (1), the verb sāwā is used
causatively (it means ‘make equal’ and has a causer participant ‘the law’), but is
syntactically intransitive (the second, non-subject, syntactic argument with the
ted

semantic role of heme is introduced by the preposition bayna ‘between’):

(1) Sāwā (intransitive, causative): [Al-Hayat 1996]
rec

al-qānūn-u y-usāw-ī bayna al-ǯamī‘-i.

def-law-nom 3sg.m-make.equal-prs.ind between def-all-gen
cor

‘he law makes everyone equal.’

Un
 Arabic ‘labile verbs’ in form III 259

y
In contrast, Example (2) represents the non-causative (spontaneous) use of sāwā

pan
(‘be, become equal’). Unexpectedly, the verb is transitive. he two symmetrical
arguments occupy the positions of subject (ad-dinār-u ‘dinar’) and direct object

Com
(aktar-a min talāt-at-i dūlār-āt-in ‘more than three dollars’).
(2) fī-mā kān-a ad-dinār-u fī ‘ām-i 1990
in-what be.pst-3sg.m def-dinar-nom in year-gen 1990

ing
y-usāw-ī aktar-a min talāt-at-i dūlār-āt-in.
3sg.m-be.equal-prs.ind more-acc from three-f-gen dollar-pl-gen

lish
‘Based on the situation in 1990, dinar (was) equal to more than three dollars.’
[Al-Hayat 1996]

Pub
hus, I show that in Arabic there are labile verbs, yet they are not numerous and
are rather peculiar both in semantic and syntactic respects. he data under analysis
lead me to a more general question: how should we treat cases where a verb has
ins
two uses which difer in their (non)-causativity but exhibit a non-standard type of
transitivity opposition 1?
am

All Arabic examples are taken from ArabiCorpus (arabicorpus.byu.edu). his

corpus appears to represent diferent regional varieties of the literary language (for
enj

instance, the list of newspapers includes Al-Watan (Kuwait), Al-Ahram (Egypt)

and Al-Hayat (diferent countries), but no dialectal data seem to be involved in the
nB

present analysis. I did my best to take most examples from Arabic newspapers in
order to make them comparable to each other. Ater each example the part of the
corpus (most oten, the name of the newspaper) is given.
Joh

Since ArabiCorpus does not have a convenient mechanism to search for all uses
of a lexeme, I mainly searched for several forms of each lexeme. he set of forms
-

included 3sg f past (qārabat), 1pl past (qārabnā), 3sg m prs (yuqāribu), and 3sg
f prs (tuqāribu). My judgments concerning frequency of lexemes are mainly based
on these forms, though some other forms were also used to check my conclusions.
ofs

he paper is organized as follows. In Section 2, a brief sketch of theoretical and

typological approaches to lability is given. Section 3 discusses some general features
pro

of transitivity and voice marking in Arabic. Section 4 is central for the paper: here
labile verbs of form III are analyzed and their typological particularity is shown. In
Section 5, I consider some cases of lability in Arabic outside form III. In Section 6,
ted

some typological parallels to the Arabic situation are addressed. In the Conclusions
section, possible theoretical implications of the Arabic data are discussed.
rec

1. Note that some authors, such as Saad (1982) do not consider Arabic verbs of symmetrical
cor

situations to be transitive at all, even if they have an accusative object. Saad points to the fact that
verbs of this sort usually do not form passives. Our corpus data in principle conirm this claim,
Un

though Google search inds some passive forms of the verb sāwā ‘be equal’.
260 Alexander Letuchiy

y
2. Deinition and distribution of labile verbs

pan
2.1 he state of research

Com
A labile verb is a verb which is able to be used transitively or intransitively without
a morphological change. Among these verbs, the A-labile and P-labile types are
distinguished.

ing
Labile (sometimes also called ambitransitive) verbs have been overlooked by
typologists for a long time. Although the deinition of labile verbs was irst pro-

lish
posed in 1940, by Jakovlev and Ashxamaf, linguistics lacked special typological or
descriptive studies dedicated to lability.
here were only some works, such as Nedjalkov (1969) where the typological

Pub
distribution of lability was analyzed. Nedjalkov, and later Haspelmath (1993), re-
gard lability as a type of formal correlation between a transitive and an intransi-
tive verb denoting one situation or semantically similar situations. For instance, in
ins
English, both the spontaneous and the externally-caused variants of the situation
‘break’ are denoted by the same lexeme break with no formal change in the verb
am

form (cf. he cup broke vs. I broke the cup).

Along with lability, other grammatical and lexical types of transitive / intran-
enj

sitive correlation exist. Here belong the causative type, the anticausative type, sup-
pletion, and equipollent oppositions. Nichols et al. (2004) propose one more type
nB

which they consider to be autonomous from the others, namely, the type with an
auxiliary change.
Joh

In the last two decades lability has increasingly become a subject of typological
research, as seen, for instance, in Haspelmath (1993), Comrie (2006), Creissels
(2009), Letuchiy (2010, 2013), Kulikov & Lavidas (2014) (the sole volume dedicated
-

to the typological aspects of lability) and others in which the authors focus on the
cross-linguistic perspective of lability studies.
ofs

Ljutikova (2002) and Letuchiy (2006) propose a semantic classiication of labile

verbs including reciprocal, relexive, anticausative (distinguished by Ljutikova) as
well as conversive and passive types (which were added to Ljutikova’s classiication
pro

by Letuchiy). In the present article, I only discuss anticausative labile verbs, which
are the only ones considered in detail by Nedjalkov (1969), Haspelmath (1993),
ted

and Comrie (2006). In this type, the transitive use of labile verbs denotes an ex-
ternally caused event, such as I broke the cup, while the intransitive use denotes
rec

an event which is conceptualized as spontaneous, having no external causer, as in

he cup broke.
cor

he main regions that exhibit labile alternations are Europe (where anticausative
lability of the break (transitive) / break (intransitive) type is observed), and Africa
Un

(the only area where passive lability is constantly found in various geographical
 Arabic ‘labile verbs’ in form III 261

y
domains and language groups, as, for example, in Songhay taka ‘bury / be buried’

pan
(Nilo-Sakharan, see Prost 1956; Rožanskij 1991; Galiamina 2005).

Com
2.2 A-lability vs. P-lability

Another opposition, which is relevant for the classification of labile verbs, is

ing
A-lability vs. P-lability.
A-labile verbs are those lexemes which have an Agent argument in both uses,
transitive and intransitive. his can be illustrated by Russian jest’ ‘eat (something) /

lish
eat (it does not matter what)’. In (3), the verb is intransitive, while in (4), it has an
accusatively-marked Patient mjaso ‘meat’:

Pub
(3) Ja je-m .
I.nom eat-prs.1sg
‘I am eating.’
ins
(4) Ja je-m mjas-o.
am

I.nom eat-prs.1sg meat-sg.acc

‘I am eating meat.’
enj

In the P-labile type, the verb has a Patient in both uses. In the transitive use, as in
(5), an Agent is usually present, while in the intransitive use, represented in (6), the
nB

verb tends to be monovalent (with only the Patient being overtly expressed). he
English boil exempliies this type:
Joh

(5) John is boiling water (an Agent and a Patient).

(6) he water is boiling, we can make some tea! (only a Patient).
-

In a particular language, either A- or P-lability can be the main (the most fre-
quent and productive) type of variation. For instance, Drossard (1998) claims that
ofs

A-lability outranks P-lability in accusative languages, while in ergative languages,

it is the opposite way round, and P-lability is more productive than A-lability.
pro

his claim calls for a more precise formulation, especially for P-lability. Both
ergative and accusative languages difer signiicantly in what concerns the num-
ber of P-labile verbs. Of course, in some ergative languages, P-lability is really
ted

frequent (Avar, Adyghe, Warekena), and in some accusative languages it is infre-

quent (Russian, Ava Pit). However, there are also accusative languages with many
rec

P-labile verbs (French, German) and ergative languages where P-lability is marginal
(Lezgian, Tsez).
cor

Letuchiy (2010) claims that accusative languages, contrary to ergative ones,

typically do not have P-labile verbs with strong semantic transitivity, like English
Un
262 Alexander Letuchiy

y
break or tear. Such is the case in Arabic, where this group of situations is not cov-

pan
ered by labile verbs.
Creissels (2014) demonstrates that the very notion of P-lability is much harder

Com
to deine in ergative than in accusative languages. In ergative languages, an intran-
sitive use of a labile verb is virtually indistinguishable from the use of a strictly
transitive verb, in which the object is omitted.

ing
2.3 Lability and syntactic transitivity

lish
When dealing with lability, linguists usually consider the relation between lability
and syntactic transitivity. he question is of real signiicance: the fact that a verb has

Pub
a causative and a non-causative use does not actually reveal much about syntactic
transitivity of the former and the latter use.
Typically, a labile verb has a transitive causative use and an intransitive
ins
non-causative one. However, Letuchiy (2013) shows that for some verbs, the vari-
ation of argument structure and semantics is not accompanied by transitivity var-
am

iation. For instance, the Turkish verb bašlamak ‘begin’ has a spontaneous use, as
in (7a), and a causative one, illustrated by (7b). Yet, neither of them is syntactically
enj

transitive. In the causative use the verb has a dative indirect object (lit. ‘he teacher
begins to a lesson’).
nB

(7) a. Ders bašl-ıyor.

lesson begin-prs
Joh

‘he lesson is beginning.’

b. öğretmen ders-e bašl-ıyor. [Scheka 1999].
teacher lesson-dat begin-prs
-

‘he teacher is beginning the lesson.’

See also Examples (1) and (2) for a similar case in Arabic.
ofs

In what follows, I will call transitive causative and intransitive non-causative

uses of labile verbs ‘canonical patterns’, since the causative meaning is usually asso-
pro

ciated to transitive verbs and the absence of causation (the spontaneous meaning)
with intransitive ones. he other two patterns, namely, intransitive causative (as
ted

in (1) and (7b)) and transitive non-causative uses (as in (2)) will be termed below
‘non-canonical patterns’.
he focus of this paper is the non-canonical cases. I will show that in Arabic
rec

form III, many verbs which are semantically similar to conventionally labile ones
do not conform to the canonical pattern: either their causative uses are intransitive,
cor

or their spontaneous (non-causative) uses are transitive.

Un
 Arabic ‘labile verbs’ in form III 263

y
In the following section, I will briely describe some relevant features of Arabic

pan
verbal system and transitivity marking. hese remarks will be important for the
discussion of labile verbs in Section 4.

Com
3. Arabic verb system and transitivity

ing
he Arabic verb system has some properties which seem to be closely related to
issues of lability. Here I only mention some of them, namely, (1) existence of nu-

lish
merous patterns of valency change marking; (2) syntactic indeiniteness of most
verb patterns, i.e., their ability to contain both transitive and intransitive verbs.

Pub
3.1 Verbal patterns ins
It is widely known that Arabic has a large system of derivational patterns, also
called forms or stems. Few of them are specialized for expression of transitivity and
am

valency, most stems can also have aspectual, modal readings or some idiosyncratic
meanings. However, almost all of these patterns impact valency and transitivity –
enj

at least with some roots. Some of the meanings related to transitivity and valency
are listed below (see Blachère 1958: 69–79; Blachère & Gaudefroy-Demombynes
nB

1952: 49–76; Grande 1998: 121–141 for the complete list of meanings):

II causative;
Joh

III applicative, conative, reciprocal;

IV causative;
V anticausative, relexive (of II);
-

VI intransitive variant of III, reciprocal;

VII passive, anticausative;
ofs

VIII anticausative, relexive;

IX anticausative / non-causative variant of form IV (with roots meaning colour,
pro

corporal defects and similar things);

X relexive.

At the same time, many of the patterns can also have uses which are not related
ted

to valency and transitivity. Some examples from [Grande 1998] are given below:
rec

II intensive action
IV intensive action
cor

V multiplicative
VI multiplicative, intensive
Un
264 Alexander Letuchiy

y
hus, the crucial property of the Arabic system is that, in spite of the richness of its

pan
voice and valency change system, it has few dedicated means of marking valency
change. 2

Com
3.2 Syntactic indeiniteness

ing
A notable peculiarity of the Arabic verbal system is its property that can be called
‘syntactic indeiniteness’. By this term I mean that nearly all morphological patterns
include verbs of diferent syntactic classes (transitive and intransitive). Among the

lish
ten frequent patterns, only VII and IX form an exception – these patterns feature
intransitive verbs exclusively.

Pub
For instance, patterns II and IV, which often form transitive verbs (II
qaddama ‘present’, IV ’alqā ‘throw, give (a lecture)’, etc.), also contain intransitive
verbs with the intensive meaning (e.g. II fallasa ‘become a bankrupt’, IV ’aǯma‘a
ins
‘gather (oneself, intransitive)’) (see Fassi Fehri 2012 for details). Patterns VIII and
X not only form intransitive relexives and anticausatives (VIII ightasala ‘wash’,
am

X istayqaẓa ‘wake up’), but also transitive verbs with the relexive benefactive or
relexive causative reading. Or, they may have the same reading as that of form I,
enj

as form VIII verb iktašafa ‘discover’ (transitive or with the preposition ‘an) or X
istašāra ‘ask for an advice’.
nB

Typologically, this property of Arabic system is not at all widespread. For in-
stance, in Turkic languages, the core system of four valency change markers (re-
lexive, reciprocal, anticausative / passive and causative) is observed (cf. Khakas -n
Joh

‘relexive’, -s ‘reciprocal’, -tyr / -t and other allomorphs ‘causative’ and -l ‘anticaus-

ative, passive’). However, only the relexive and reciprocal markers can form both
-

transitive and intransitive verbs. he two other markers – the causative and the an-
ticausative / passive one – derive only transitive and intransitive verbs, respectively.
On the other hand, many languages behave similarly to Arabic. For instance, in
ofs

Adyghe (West Caucasian) relexive, reciprocal, and many applicative preixes can
produce transitive as well as intransitive forms. In any case, the property of syntactic
pro

indeiniteness manifests itself very clearly in Arabic – it afects even the causative
marker which oten is unable to form intransitive verbs cross-linguistically.
In the next section, I will show that the non-standard type of variation in
ted

form III is related to the syntactic indeiniteness of this form. he key property
which permits the occurrence of variation is that form III includes both transitive
rec

and intransitive verbs, though the latter are less frequent than the former.
cor

2. One of the exceptions is formed by the passive formation marked by the vowel change (kataba
Un

‘write’ – kutiba ‘be written’) which is beyond the scope of the present article.
 Arabic ‘labile verbs’ in form III 265

y
3.3 Form III and non-standard transitivity

pan
In many traditional works, such as Blachère (1958) and Grande (1998), the third

Com
form is said to contain mainly or only transitive verbs. In some respects, the claim
about the transitivity of form III appears to be true: verbs of form III sometimes
denote situations which do not tend to be expressed by transitive verbs cross-lin-
guistically, and follow the transitive pattern in many of these cases. At the same

ing
time, we should take into account the semantics of the base verb and the particular
use of form III.

lish
For instance, the comitative applicative use of form III, illustrated by (8) (ǯālasa
‘sit with’ from ǯalasa ‘sit’) makes the initial verb transitive, which agrees with the
deinition given by Blachère (1958: 69–79): “A la IIIme forme, le verbe exprime la

Pub
direction dans laquelle s’accomplit l’action”. he same is true for the reciprocal use,
illustrated by (9): the transitive verb kataba ‘write’ remains transitive in form III
(kātaba ‘write to each other’), the object position now being occupied by one of
ins
the reciprocal participants.
am

(8) fa qad kān-a j-uǯālis-u aṣdiqā-’a-hu

and already be.pst-3sg.m 3sg.m-sit.iii-prs.ind friend.pl-acc-3sg.m
enj

min as-sijjāsij:-īna. [Al-Hayat 1996]

from def-politician-gen.pl
nB

‘And he had sat with his friends, politicians.’

(9) Lays-a al-ǯinirāl-u man y-ukātib-u-hu.
Joh

not.be.pst-3sg.m def-general-nom who 3sg.m-write-prs.ind-3sg.m

‘It was not the general who changed letters with him.’ [Al-Hayat 1996]
-

See Arkhipov (2009) for possible patterns of marking arguments of the comitative
construction and Nedjalkov & Geniušienė (2007) for reciprocals. We can say that
cross-linguistically, the comitative participant (= the person or entity that partici-
ofs

pates in the situation together with another participant) and the second participant
of reciprocal situations tend to be marked by a non-accusative case, for instance,
pro

instrumental, dative etc., in grammatically-marked reciprocal and comitative

constructions.
ted

(10) a. Ivan pocelova-l-Ø Daš-u.

Ivan.nom.sg kiss(pf)-pst-sg.m Dasha-acc.sg
rec

‘Ivan kissed Dasha.’

b. Ivan i Daš-a pocelova-l-i-s’.
cor

Ivan.nom.sg and Dasha-nom.sg kiss(pf)-pst-pl-refl

‘Ivan and Dasha kissed each other.’
Un
266 Alexander Letuchiy

y
he Arabic situation can be accounted for if the uses of form III in (8) and (9) are

pan
regarded as instances of applicative. Peterson (2007) shows that among the uses of
applicative markers in the world’s languages there is a comitative use too. Reciprocal

Com
constructions are not usually regarded as a type of applicatives, yet, this can be done
for transitive verbs like kātaba.
At the same time, form III has some less predictable uses. For instance, this pat-
tern is used to form denominal verbs. hese cases demonstrate signiicant variance

ing
in their transitivity features. For instance, sāfara ‘travel’ is intransitive:

lish
(11) Sāfar-tu ilā al-qāhirat-i fa istaqbal-a-nī ṣadīq-ī.
travel.pst-1sg to def-Cairo-gen and meet-3sg.m-1sg friend-1sg
‘I went to Cairo, and my friend met me there.’ [Al-Hayat 1996]

Pub
It may be the case that many intransitive verbs of form III are derived from nouns,
and not from verbs, and denote a property or an event related to this noun (ver-
ins
bes qualitatifs or dénominatifs in terms of Blachère & Gaudefroy-Demombynes
1952: 56, 59). Perhaps their intransitivity results from the fact that the relations be-
am

tween the verb and the base noun demonstrate cross-linguistically a high degree of
variation and are not necessary reducible to well-described notions like ‘applicative’,
enj

‘causative’, and so on. he meaning of the verb derived from the noun is to a high
degree inluenced by the meaning of the noun, and not by syntactic restrictions.
nB

4. Lability in form III

Joh

In this section, I consider some labile verbs attested in form III. hey mainly be-
long to the group of ‘symmetrical situations’, 3 or inherent reciprocals in terms of
-

Kemmer (1991). However, a more precise deinition is necessary since symmetrical

situations like ‘kiss’ are not designated by labile verbs in Arabic. It seems more
ofs

plausible that the group of labile verbs in form III denotes symmetrical states: it
mainly includes verbs like ‘be equal’, ‘be / become close’, ‘reconcile’, where the verb
pro

denotes a state or a starting point of the state, and the state of the participant A
with respect to B is indistinguishable from the state of B with respect to A. Some
verbs like ṣālaħa ‘reconcile’ are dynamic, but they share with symmetrical states
ted

the feature of indistinguishability of the two sub-situations: if A reconciled with B,

it is automatically true that B reconciled with A. In contrast, situations like ‘kiss’ or
rec

3. By this term, I mean those situations which have two or several participants which take part
cor

in the situation in the same way. For instance, the situation ‘buy’ is not symmetrical, because the
role of buyer difers from the role of seller. he situation ‘be equal’ is symmetrical, because the
Un

two entities which are equal has the same semantic role in the situation.
 Arabic ‘labile verbs’ in form III 267

y
‘hug’ are purely dynamic and do not contain any state component in their meaning.

pan
Moreover, A’s actions towards B are distinguishable from B’s actions towards A (it
is possible that B did not kiss A, even if A kissed B).

Com
he peculiarity of the group of form III labile verbs is that many of them have
more than two syntactic patterns. Both their causative and non-causative use can
be either transitive or intransitive.
his fact comes in sharp contrast with the deinition of (anticausative) labile

ing
verbs which almost always presuppose a transitive causative and an intransitive
non-causative use. In many languages, cases of the Arabic type are rare, though,

lish
as I have shown above (see Examples (7a) and (7b)), they are deinitely not unique
cross-linguistically.

Pub
As I will demonstrate below, verbs in the group under analysis difer from
each other by the range and frequency of syntactic patterns. For example, sāwā ‘be
equal, make equal’ is mainly used in the two non-canonical patterns: the transitive
ins
non-causative and the intransitive causative ones. Fāraqa ‘go away, leave; divide’ is
primarily found in the non-canonical transitive non-causative use and sometimes
am

in the intransitive causative use. Finally, qāraba ‘make / get closer’ has all the four
theoretically possible patterns: the transitive and intransitive causative construc-
enj

tions and the transitive and intransitive non-causative constructions.

nB

4.1 sāwā

he verb sāwā ‘be equal; make equal’ is mainly used in two patterns, both of which
Joh

are non-canonical in the terms used above (see 2.3). he irst is a causative intransi-
tive pattern, as in (12) (the object is not in the accusative case, but is introduced by
-

the preposition bayna ‘between’); 4 the second is a non-causative transitive pattern,

as in (13).
ofs

(12) al-qānūn-u y-usāw-ī bayna al-ǯamī‘-i.

def-law-nom 3sg.m-make.equal-prs.ind between def-all-gen
‘he law makes everyone equal.’ [Al-Hayat 1996]
pro
ted

4. An important question, posed by the anonymous reviewer, concerns the degree of obligatori-
ness of bayna ‘between’ in examples like (12). I do not possess any information for a wide lexical
rec

sample, yet, it seems that for some verbs in some examples, bayna is obligatory. For instance, no
example parallel to (12) and lacking bayna is found in the corpus. Another reviewer’s remark
has to do with stylistic attribution of bayna in the ‘symmetrical’ use. It seems that the occurrence
cor

of bayna-construction in newspapers of various regions prevent us from classifying this use as

purely colloquial, though no analysis of its occurrence in iction and other types of texts has been
Un

carried out.
268 Alexander Letuchiy

y
(13) fī-mā kān-a ad-dinār-u fī ‘ām-i 1990

pan
in-what be.pst-3sg.m def-dinar-nom in year-gen 1990
y-usāw-ī aktar-a min talāt-at-i dūlār-āt-in.

Com
3sg.m-be.equal-prs.ind more-acc from three-f-gen dollar-pl-gen
‘Based on the situation in 1990, dinar (was) equal to more than three dollars.’
[Al-Hayat 1996]

ing
Note that no examples of the intransitive non-causative pattern, for instance, with
the preposition li ‘to, for’ or ma‘a ‘with’ are found in any newspaper. 5

lish
he transitive causative pattern is attested in newspapers, but only once and in
a context that seems semantically obscure:

Pub
(14) tumma iktašaf-nā ’anna lawn-a wujūh-i-nā abyaḍ-u
aterwards ind.out-1pl that color-acc face.pl-gen-1pl white.m-nom
min ’an y-usāw-ī-nā bi as-su’ā d-i as-sā’id-i.
ins
from to 3sg.m-make.equal-prs. ind-1pl with p.n.
‘hen we found out that the color of our faces is white, too white to make us
am

equal to Suad Said (?).’ [Al-Ahram 1999]

his means that the parameters of transitivity and causativity interact in a peculiar
enj

way in the uses of the verb sāwā, as well as other verbs analyzed below.
As early as in (1980), Hopper and hompson published a famous article where
nB

they propose that syntactic (in)transitivity of verbs interacts directly with their
semantics. Among the semantic parameters which are taken to be related to tran-
Joh

sitivity is the agentivity of the subject participant, the patientivity of the object,
and the number of participants. All of these parameters predict that the causative
variant of the situation ‘make equal’ must be more transitive than the non-causative
-

variant ‘be equal’: the former is (or at least can be) dynamic, its subject is Causer,
though not obligatorily agent, while the latter is usually a static event, and its subject
ofs

is Patient or heme, as well as the object.

he reason of non-standard behavior of the Arabic verb sāwā can be that the
pro

transitive verb has two symmetrical objects, cf. ‘X makes Y equal to Z / X makes
Y and Z equal’, where Y and Z are symmetrical. he existence of the pattern with
bayna allows a native speaker to express the symmetricity in a direct way: ‘X makes
ted

equal between Y and Z’, ‘X makes the relation between Y and Z symmetrical.’
rec
cor

5. Transitive patterns were searched in the following way: in the beginning, the irst 100 exam-
ples of each verb were analyzed, second, all examples with short object pronouns (e.g. yusāwi-hi
Un

‘it is equal to it’) were analyzed separately.

 Arabic ‘labile verbs’ in form III 269

y
4.2 mātala

pan
For the verb mātala ‘make / consider similar; be similar’, the situation is a bit difer-

Com
ent. he intransitive causative pattern with bayna ‘between’ is much less widespread
than with sāwā. For instance, it occurs only twice in the whole ArabiCorpus (both
examples are from the Al-Hayat newspaper), cf. (15) and (16):

ing
(15) al-‘unwān-u alladī y-umātil-u bayna
def-name-nom which.sg.m 3sg.m-equalize-prs.ind between

lish
ibn-i khaldūn-i wa al-ǯābirī. [Al-Hayat 1997]
son-gen Khaldun-gen and def-Jabiri
‘he name which equalizes Ibn-Khaldun and Al-Jabiri.’

Pub
(16) y-umātil-u bayna ra’y-i al-fawṭi wa
3sg.m-make.similar-prs.ind between view-gen def-Fawti.gen and
[Al-Hayat 1996]
ins
ra’y-i bašar ibn-i mu‘tamar
view-gen Bashar son-gen Mutamar
am

‘He considers Al-Fawti’s and Bashar Ibn Mutamar’s views similar.’

In general, the causative use of the verb is rare, both with bayna and with a direct
enj

object (the latter almost never occurs in texts). Furthermore, my search in the irst
examples of this verb did not show any example of the transitive causative pattern.
nB

In contrast, mātala is oten used with the transitive pattern in the non-causative
reading. For this verb, the prevalence of the ‘non-canonical’ pattern (syntactically
Joh

transitive but semantically non-causative) is evident.

(17) al-irḍa‘-u aṣ-ṣina‘iyy-u lā y-umkin-u
def-nursing-nom def-artiicial-nom not 3sg.m-be.possible-prs.ind
-

’an y-umātil-u aṭ-ṭabī‘iyy-a [hawra]

to 3sg.m-be.similar-prs.ind def-natural-acc
ofs

‘he artiicial nursing cannot be similar to the natural one.’

pro

4.3 ṣālaħa
ted

he verb ṣālaħa ‘reconcile; become reconciled’ usually occurs in the two non-ca-
nonical patterns: either in the causative intransitive pattern, with the preposition
rec

bayna ‘between’ (18), or in the non-causative transitive pattern (19).

cor
Un
270 Alexander Letuchiy

y
(18) al-muǯtama‘-u al-madaniyy-u alladī y-uṣāliħ-u

pan
def-society-nom def-citizen-nom which.m.sg 3sg.m-reconcile-prs.ind
bayna ad-dīn-i wa al-‘ilm-i wa al-māl-i.

Com
between def-religion-gen and def-science-gen and def-money-gen
‘Civil society that reconciles the religion, the science and the money with each
other.’ [Al-Hayat 1997]
(19) hadā al-mutaqqaf-u y-ukhaṭim-u as-sulṭat-a

ing
this.sg.m def-intellectual-nom 3sg.m-conlict-prs.ind def-power-acc
fī aṣ-ṣabāħ-i wa y-uṣālih-u-hā fī

lish
in def-morning-gen and 3sg.m-reconcile-prs.ind-3sg.f in
al-masā’-i. [Al-Hayat 1996]

Pub
def-evening-gen
‘his intellectual conlicts with the power in the morning and becomes recon-
ciled with it in the evening (i.e. changes his political views very oten).’
ins
he transitive causative pattern also occurs, as in (20), though only in two examples
am

in the entire corpus. he intransitive non-causative pattern (21) is attested only once
and only from an older text.
enj

(20) ’anti allatī ṣālah-at-nī ma‘a allah-i. [Remembrance]

you.sg.f which.sg.f reconcile-3sg.f-1sg with God-gen
nB

‘It was you who reconciled me with God.’

(21) la‘alla-nā n-uṣālih-Ø ‘alā šay’-in. [Aghani]
Joh

part-1pl 1pl-reconcile-subj on thing-gen

‘Let’s reconcile (= come to one opinion) with respect to some questions.’
-

4.4 fāraqa
ofs

he verb fāraqa ‘divide; leave’ is mainly found in the corpus in one non-canonical
pattern, speciically the transitive non-causative (22). he intransitive causative
pro

use (with the preposition bayna ‘between’) is found only once, as in (23), while the
other two patterns are not found at all.
ted
rec
cor
Un
 Arabic ‘labile verbs’ in form III 271

y
(22) Wa ad-du‘ā’-u y-ufāriq-u-hum Wa hādā

pan
and def-prayer-nom 3sg.m-leave-prs.ind-3pl.m and this.sg.m
an-naṣr-u li al-muqa:wamat-i. [hawra] 6

Com
def-victory-nom for def-opposition-gen
‘And the prayer leaves them, and this is the victory for the opposition.’
(23) Wa hadā mā y-ufāriq-u bayna al-’īmān-i

ing
and this.sg.m what 3sg.m-divide-prs.ind between def-belief-gen
ad-dīniyy-i Wa al-’īmān-i al-buṭūliyy-i. [Al-Hayat 1996]

lish
def-religious-gen and def-belief-gen def-heroic-gen
‘his is what distinguishes religious belief from heroic belief.’

Pub
he meaning of the verb in the causative and the non-causative patterns is not
identical. he causative meaning is ‘distinguish’, while the non-causative sense is
‘leave.’ his semantic variant does not have a causative correlate, such as ‘divide (the
ins
people), make smb. leave smb.’
am

4.5 qāraba
enj

he verb qāraba ‘make / get closer’ behaves like the majority of the verbs analyzed
here in that it has an intransitive causative use, as in (24)–(25), and a transitive
nB

non-causative use, as in (26).

(24) Wa inna li-īṭāliya ayḍan waǯh-an’ ūrūbiyy-an šarqiyy-an aw
Joh

and part to-Italy also face-acc european-acc oriental-acc or

mā y-uqārib-u bayna-hā. [Al-Hayat 1996]
-

what 3sg.m-make.close-prs.ind between-3sg.f

‘In fact, Italy also has a mixed face (european and oriental) or something which
combines the two of them.’
ofs

(25) Wa y-uqārib-u bayna mu‘āmalat-i-ha hunā Wa

and 3sg.m-make. close-prs.ind process-gen-3sg.f here and
pro

al-mu‘āmalat-i allatī kāna-t al-ħālat-a aṣ-ṣiħħiyy-at-a.

def-process-gen which.sg.f be-3sg.f def-case-acc def-true-f-acc
ted

‘And he compares (lit. makes close) the process (which took place) here and
the process which takes place in the correct case.’ [Ghad 2001]
rec

6. Example (22) can also be understood as ‘he prayer divides them, makes them diferent’.
cor

However, this does not change the general tendency that the transitive use of fāraqa is non-caus-
ative. Note, however, that more examples for the pattern with bayna ‘between’ can be found on
Un

Google.
272 Alexander Letuchiy

y
(26) …fī bank-i ar-riyāD-i alladī y-uqārib-u-hu fī

pan
in bank-gen Riyadh-gen which.sg.m 3sg.m-be.close-prs.ind in
ar-ra’asmāl-i [Al-Hayat 1997]

Com
def-capital-gen
‘… In the Riyadh bank which has a similar capital.’

However, this lexeme shows more variation than the verbs previously analyzed: it

ing
evidences both an intransitive non-causative use illustrated in (27) and a transitive
causative use, cf. (28):

lish
(27) …mā y-uqārib-u min 300 milyūn-i dīnār-in
what 3sg.m-be.close-prs.ind from 300 million-gen dinar-gen

Pub
kuwaytiyy-in. [Al-Ahram 1999]
Kuwait.adj-gen
‘… what is close to 300 million Kuwait dinars.’
ins
(28) Wa lā y-uqārib-u-hā min munṭalaq-in ’anna-hā
and not 3sg.m-make.close-prs.ind- 3sg.f from start-gen that-3sg.f
am

al-furṣat-u al-akhirat-u [Ghad 2001]

enj

def-chance-nom def-last-nom
‘And the fact that it is the last chance does not make her closer to the starting
nB

point.’

Note that the last two patterns are also found with two other verbs of the same
Joh

root: the latter (transitive causative) is absolutely equivalent to the one exhibited
by qarraba of form II, while the former (intransitive non-causative) is oten found
with taqarraba in form V or iqtaraba in form VIII. Moreover, the existence of these
-

two forms which are unambiguously transitive and intransitive, respectively, can
result in contamination of the transitive causative and intransitive non-causative
uses in the verb of form III.
ofs
pro

4.6 ṭābaqa

he verb ṭābaqa means ‘be adequate, be compatible, equalize, compare’. It is found

ted

in the non-canonical causative use (with the preposition bayna), as in (29) and
the non-canonical non-causative use, as in (30). Note that the non-causative use is
rec

much more frequent than the causative one.

cor
Un
 Arabic ‘labile verbs’ in form III 273

y
(29) …alladī y-uṭābiq-u bayna ad-dawlat-i wa

pan
which.sg.m 3sg.m-make.equal-prs.ind between def-state-gen and
al-ummat-i. [Al-Hayat 1997]

Com
def-community-gen
‘…Which makes the state and the community compatible with each other.’
(30) wa la y-abdū ’anna hadā al-kalām-a

ing
and not 3sg.m-seem.prs.ind that this.m def-speech-acc
y-uṭābiq-u al-wāqi‘-a. [Al-Hayat 1996]

lish
3sg.m-be.equal-prs.ind def-fact-acc
‘It does not seem that this claim relects the reality.’

Pub
he canonical non-causative use does not occur. he canonical causative use is
rarely found, though there are some uses with the preposition ma‘a ‘with’. 7
ins
4.7 wāzana
am

he verb wāzana ‘compare, consider equal; be equal, weigh equally’ is almost ex-
clusively found in one syntactic pattern: namely, the causative intransitive use with
enj

bayna, illustrated by Example (31). he non-causative transitive use is rarely found

(32), while the causative transitive and the non-causative intransitive uses are each
nB

found in only one example in the corpus.

(31) y-uwāzin-u bayna ħāǯ-āt-i-hi wa ħāǯ-āt-i
Joh

3sg.m-compare-prs.ind between need-pl-gen-3sg.m and need-pl-gen

al-ākhar-īna. [Al-Hayat 1996]
-

def-other-gen
‘He compares (makes equal) his needs and other’s needs.’
(32) lā y-ataṣādam-u
ofs

ma‘a-hu bal y-uwāzin-u-hu

not 3sg.m-confront-prs.ind with-3sg.m but 3sg.m-be.equal-prs.ind-3sg.m
‘He does not confront it, but balances it.’
pro
ted
rec

7. he canonical causative use with the preposition bi ‘with’ is rather frequent. However, these
cor

uses with this preposition are let beyond the scope of the article, since some linguists regard
this preposition as an analytical marker of causative, based on examples like jā’a ‘come’ vs. jā’a bi
Un

‘bring’ (see, for instance, Khrakovkij 2011).

274 Alexander Letuchiy

y
4.8 Intransitive structures with bayna: What is the functional motivation?

pan
In the linguistic evidence presented above the pattern with bayna ‘between’ consti-

Com
tutes the most striking part of the data. While a non-causative verb can be transitive
or intransitive (both variants are found across languages, and their distribution
depends crucially on verbal semantics and the nature of arguments, see Næss (2007)
for detailed analysis), causative verbs typologically tend to be transitive.

ing
However, this behavior agrees well with the generalizations drawn by Hopper
and hompson (1980) and Tsunoda (1985). Note that the prototypical transitive

lish
situation which is served by transitive verbs in the world’s languages is a bivalent
situation. he trivalent situation is always subject to complicated processes which
result from concurrence between the two objects for the status of the direct object.

Pub
Recall, for instance, that among the languages of the world there are both those
in which the addressee / recipient of the trivalent verb behaves as a privileged object
(= primary object construction) and those where the patient / theme is more likely
ins
to be the highest object (= direct object construction), see Malchukov et al. (eds)
2010 for details.
am

With trivalent causative verbs of symmetrical relations, just as sāwā, mātala

and others discussed throughout this paper, the situation is even more complicat-
enj

ed. Not only is there a choice between two object arguments pretending to be a
privileged object, but also their semantic properties are completely identical. If the
nB

agent / causer A makes B and C symmetrical to each other (e.g., makes them equal
or close to each other), there is no way to prove that either B or C is semantically
Joh

a more canonical direct object.

If we formulate the situation using the mechanism of competing motivations,
proposed in the framework of the Optimality heory, the bayna-pattern, which
-

is observed in (12) and (15), and the canonical transitive causative pattern repre-
sented, for instance, in (14) and (28), each observe one constraint but violate the
ofs

other. he constraints are:

1. ‘follow the prototypical causative pattern’ (i.e., a causative verb must be

pro

transitive);
2. ‘mark the arguments accordingly to their semantic properties’ (i.e., arguments
with identical semantic properties must be identically marked).
ted

he canonical transitive pattern follows constraint 1 because transitive behavior

rec

is normal for causative verbs. However, it violates constraint 2 because, of the two
identical arguments, one becomes a direct object, and another one is designated
cor

by a PP. In contrast, the bayna-pattern does not follow the causative prototype
(constraint 1), but marks the identical arguments by identical means (constraint 2).
Un
 Arabic ‘labile verbs’ in form III 275

y
Note that in some languages, bayna-like patterns are prohibited for verbs. For

pan
instance, in Russian, a bayna-like preposition meždu ‘between’ can only introduce
symmetrical arguments of nouns, but not of transitive verbs:

Com
(33) Dal’tonik-i ne različaj-ut sin-ij i
daltonics-pl.nom not distinguish-prs.3pl blue-m.sg.acc and
krasn-yj. / *meždu sin-im i krasn-ym.

ing
red-m.sg.acc between blue-m.sg.ins and red-m.sg.ins
‘Daltonics do not distinguish blue and read.’

lish
(34) Meždu et-imi cvet-ami jest’ raznic-a.
between this-pl.ins color-pl.ins be.prs.3sg diference-sg.nom

Pub
‘here is a diference between these colors.’

Languages like Russian rank the causative pattern constraint higher than the iden-
tical marking constraint. In such languages, in general, one of the two objects of
ins
trivalent verbs is usually marked as a direct object, even when neither of them is a
prototypical patient.
am
enj

4.9 Variation in form III: Summary

To sum up, in the previous sections I showed that some Arabic verbs like ‘be equal’
nB

and ‘be close’ of form III show syntactic and semantic variation. All these verbs
belong to the class of symmetrical states, in which the state of the participant A is
Joh

indistinguishable from the state or actions of another participant, B. In other words,

not only are the situations denoted by the verbs under analysis always or typically
symmetrical, but also they belong (at least in some uses) to the state class (cf. ‘be
-

equal’, a situation which presupposes no dynamic changes).

There are two features which distinguish the Arabic lexemes analyzed in
ofs

Sections 4.1 through 4.7 from canonical labile verbs. First, the non-causative uses
of the verbs under analysis do not denote canonical ‘spontaneous events’ as deined
pro

by Haspelmath (1993). Haspelmath (1993) mainly analyzes verbs denoting dynam-

ic events, while the Arabic lexemes under analysis mainly denote states (in other
words, both semantics like ‘be equal’ (state) or ‘become equal’ (event) are available,
ted

e.g., for the verb sāwā ‘be / make equal’).

Second, while Nedjalkov (1969), Haspelmath (1993), and Comrie (2006) con-
rec

sider the semantic parameter of spontaneity and the syntactic parameter of tran-
sitivity to be correlated, this is not the case with the lexemes considered above.
cor

It has been demonstrated that the verbs under analysis can have both transitive
non-causative and intransitive causative uses. In Table 1, the uses of these verbs
Un

are summarized. he labels for the uses are CT (causative transitive), CI (causative
276 Alexander Letuchiy

y
intransitive, with the preposition bayna ‘between’), NT (non-causative transitive)

pan
and NI (non-causative intransitive). Most data are taken from the corpus, of course
some of uses which have not been found in the corpus can easily be found on the

Com
Internet. +? is used if the given use is found only once or twice in the corpus.

Table 1. Uses of form III verbs showing syntactic and semantic variation
verb CT CI NT NI

ing
sāwā +? + +
mātala + +

lish
ṣālaħa +? + + +?
ṭābaqa +- + +

Pub
qāraba + + + +
bā‘ada + + +
wāzana +?/– + ins + +?

Note that the uses which demonstrate non-standard correlation of syntactic features
with the meaning of the verb (transitive non-causative, intransitive causative) are
am

the ones which are found with the greatest number of verbs.
Before going further, a remark should be made here on the semantic relation
enj

between verbs of form III and their base verbs of form I. It cannot be claimed that
the third form uniformly changes the syntactic properties of the base form I verb.
nB

he semantic correspondence between form III of these verbs and their base verbs
in form I varies signiicantly even across the small group under analysis, as Table 2
Joh

shows. 8
hus, it should be concluded that, independently of the starting point of the
derivation (form I verb), when we get to the inal point (form III verb), the verb
-

can acquire lability. I propose that this fact results from the semantics of form III.
In the meaning of form III, the component of symmetricity (symmetrical situation)
ofs

is proiled. All other components, including (non)-spontaneity of the situation, are

backgrounded. (Recall that the feature of syntactic indeiniteness characterizes the
Arabic system: most verb forms contain both transitive and intransitive verbs.)
pro

his is why the resulting verb of form III can have either causative or non-causative
meaning.
ted

This is reminiscent of the language phenomena which Bybee (1995) calls

‘product-oriented schemas’, the main diference being that Bybee employs this
rec

notion for morphological and phonological processes rather than for syntax and
cor

8. At the same time, the semantic correspondence between other groups of form III verbs with
their base verbs is much more regular: for instance, there is a semantically homogenous group
Un

of comitative applicatives (cf. Example (8)).

 Arabic ‘labile verbs’ in form III 277

y
Table 2. Semantic correspondence of form III verbs with base form I verbs

pan
verb meaning verb of form I meaning
sāwā ‘make / consider / be equal’ sawiya ‘cost’; ‘be equal’

Com
mātala ‘make / consider/ be similar’ matala ‘appear, become, be similar to,
sculpture, disigure (with the
preposition bi)’
ṣālaħa ‘reconcile / be reconciled’ ṣalaħa or ṣaluħa ‘be good, religious’; ‘be plausible

ing
for smth.’
ṭābaqa ‘be equal / equalize’ no verb of form I

lish
qāraba ‘make / be close’ qaruba ‘be close’
bā‘ada ‘make / be farther’ ba‘uda ‘be far away’; ‘be incredible’
wāzana ‘compare, make / consider wazana ‘weigh (transitive causative)’

Pub
equal; be equal, weigh wazuna ‘be serious; weigh a lot’
equally’ ins
semantics. Bybee means that some morphological processes yield the same result
(for instance the same morpho(phono)logical type of the word or the syllable),
am

independently of the starting point (the base word or the syllable). his is what
takes place in Arabic, though with a rather small class of symmetrical situations.
enj

hese verbs become labile in form III independently of the syntactic features of
the base verb.
nB

Another possible solution is the following: it can be claimed that the majority of
the base verbs are intransitive, and that the derivation of form III adds a causative
Joh

component, for instance, ‘make equal,’ to the initial use. However, this analysis is
not unproblematic, because the main semantic component which distinguishes
form III of the verbs under analysis from form I is the symmetricity of the event
-

and not the causative meaning.

Finally, there is an explanation lying in the derivation domain. We could say
that the group of verbs under analysis is derived from adjectives or nouns (see
ofs

Blachère & Gaudefroys-Demombynes (1952: 56, 59, etc.), where a special class of
denominal / deadjectival verbs is considered). As stated above, deadjectival verbs
pro

oten show lability because both patterns are derived from non-verbal stems, and
the transitivity feature is neutralized. In principle, all verbs discussed in Section III
ted

have some parallel adjectival formations (e.g., sawā’un ‘equal’ for sāwā ‘be, make
equal’; ṣāliħun ‘good’ for sālaħa ‘reconcile, become reconciled). Letuchiy (2013)
shows that in many languages, deadjectival verbs tend to be labile. However, the
rec

very claim of deadjectival origin of the verbs under analysis is hard to verify. First,
Arabic deadjectival verbs contain no specialized markers of derivation, distinguish-
cor

ing them from other types of verbs. Second, the deadjectival origin of verbs like
Un
278 Alexander Letuchiy

y
‘be equal’ or ‘be close’ is not as clear semantically as, for instance, in the case of

pan
‘consider fool’ or ‘make black’.

Com
5. Labile verbs of other forms

Lability is not restricted to the form III in Arabic. However, in other patterns it

ing
seems to occur only sporadically and is not a systemic phenomenon. he two com-
mon features of the form III group, analyzed above, and other groups of labile verbs,

lish
is that (i) there are no labile verbs with a strong semantic transitivity, such as ‘verbs
of killing’ or ‘verbs of breaking’ in terms of Levin (1993); and (ii) most of these verbs

Pub
are derived, though some exceptions exist.
For instance, in form I, the verb bada’a ‘begin’ is labile. Note that alongside
the monovalent intransitive and the bivalent transitive patterns, it can also form
ins
a bivalent intransitive structure where the second argument is introduced by the
preposition bi (see Saad 1982 on this type of causatives). None of the three patterns
am

are marked as dialectal, and all of them occur in newspapers. However, the precise
distribution of the transitive and intransitive bivalent patterns remains unclear.
enj

Consider the following examples:

(35) y-abda’-u ar-ra’īs-u al-miṣriyy-u ħusnī
nB

3sg.m-begin-prs.ind def-president-nom def-egyptian-nom p.n.

mubārakun ziyārat-an ilā wāšintūn. [Al-Hayat 1997]
Joh

p.n. visit-acc to Washington

‘he president of Egypt Husni Mubarak begins his visit to Washington.’
(36) y-abda’-u
-

al-kitāb-u bi ‘arḍ-in ‘āmm-in.

3sg.m-begin-prs.ind def-book-nom with presentation-gen general-gen
‘he book begins with a general presentation.’ [Al-Hayat 1997]
ofs

(37) li n-abda’-Ø bi mawḍū‘-i ṣudūr-i al-qarār-āt-i.

let’s 1pl-begin-subj with topic-gen publishing-gen def-decision-pl-gen
pro

‘Let’s begin (to discuss) the topic of taking decisions.’ [Al-Hayat 1997]

In form V, the verb tabayyana ‘become clear; make clear, explain’ is labile. In the
ted

intransitive use the Content which becomes clear occupies the subject position (38),
however, in the transitive use the same participant is an object, the subject position
rec

being reserved for an animate Agent (or Experiencer) (39):

cor
Un
 Arabic ‘labile verbs’ in form III 279

y
(38) tabayyan-a ’anna hunāka qiwā musta‘idd-at-u li

pan
become.clear-3sg.m that there power.pl ready-f-nom for
al-wutūb-i. [Al-Ahram 1999]

Com
def-rise-gen
‘It became clear that there are powers which are ready for the rise.’
(39) fa qad tabayyan-ū-hu min ṣudūr-i aṣ-ṣaħīfat-i.

ing
and already ind.out-3pl.m-3sg.m from issue.pl-gen def-newspaper-gen
‘hey found it out from issues of the newspaper.’ [Al-Ahram 1999]

lish
Some verbs have both transitive and intransitive uses, but difer from canonically
labile verbs in that the two uses are semantically distinct from each other – in other

Pub
words, it is not only the causative meaning that distinguishes the transitive use from
the intransitive one. his phenomenon has been discussed by Daniel et al. (2012)
for Daghestanian languages. In Arabic, this may be seen, for instance, in tadāwala
ins
‘discuss, use; be in use’. While the intransitive use can be employed both with names
of money units which are in use in the particular territory, and with topics people
am

discuss, the transitive use, on the other hand, is found primarily with the meaning
‘discuss’. Of course, these two meanings are historically related: if someone discusses
enj

a topic, (s)he in a sense uses it, and this topic is transferred from one person to
another in the same sense as money makes its way from one person to another.
nB

(40) Wa tadāwal-at wasā’il-u al-’i‘lān-i ’ism-a

and discuss-3sg.f means.pl-nom def-information-gen name-acc
Joh

‘abd-i al-maǯīd-i al-kāriħ-i. [Tajdid 2002]

p.n.
‘And the press discusses the name of Abdu-l-Madzhid-al-Karih.’
-

(41) Wa khilāla al-ǯalasāt-i tadāwal-at fī-hā ’awrāq-un

and in.course.of def-session-gen discuss-3sg.f in-3sg.f sheet.pl-nom
ofs

iḍḍiyy-at-un. [Al-Ahram 1999]

silver-f-nom
pro

‘And in the course of the discussions there were discussed / used ‘silver
securities’.’
ted

As I mentioned earlier, the lability of these verbs does not have a systemic character
which is at least partially observed in the verbs of symmetrical events in form III.
rec

his is why they are not considered here in detail (see Letuchiy 2013: 302–339 for
details).
cor
Un
280 Alexander Letuchiy

y
6. Lability of derived verbs: Some parallels in other languages

pan
hroughout this article, I have emphasized that derived verbs of form III are more

Com
plausible candidates for being labile in Arabic than base verbs of form I. Although
this fact might be incidental, taking into account the small number of labile verbs in
Arabic, I consider it to be the result of a feature of the Arabic verbal system which
was termed here ‘syntactic indeiniteness.’ I argued that most verb forms can derive

ing
both transitive and intransitive verbs.
For instance, form III can be either transitive or intransitive. What is central

lish
for the behavior of the derived form is the semantic component of symmetricity.
In contrast, the presence or absence of the causative component is peripheral. his

Pub
results in the fact that form III verbs can be either causative or non-causative.
he same is applicable to some other labile verbs. For instance, form V can
either denote anticausative (in this reading it derives intransitive verbs) or middle
ins
(relexive benefactive) which yields transitive verbs. he verb tabayyana combines
the two readings: it is both an anticausative of bayyana ‘ind out, explain’ (this re-
am

sults in its intransitive meaning) and a relexive benefactive of the same verb (this
is why the same lexeme has a transitive meaning ‘ind out, explain to oneself ’).
enj

hough the syntactic indeiniteness is not characteristic of European languages,

the tendency observed in Arabic is relected in Europe. In many European lan-
guages, there are derivational patterns which oten yield labile verbs. Here belongs,
nB

for instance, the pattern observed in French, where some verbs of the ir-type are
derived from adjectives:
Joh

(42) French (work with native speakers):

noir ‘black’ – noircir ‘become / make black’;
-

blanc ‘white’ – blanchir ‘become / make white’

he German pattern of deriving verbs from adjectives and nouns by means of the
ofs

preix ver- also yields some labile verbs:

German (work with native speakers):
pro

(43) verkohlen ‘become coal, be carbonized; carbonize’.

Kehayov (in press) conirms the same tendency for Estonian where derived verbs
ted

manifest lability more oten than base verbs.

rec
cor
Un
 Arabic ‘labile verbs’ in form III 281

y
7. Conclusions

pan
In the present paper, I focused on a special group of Arabic labile verbs, namely,

Com
verbs of form III. his group is semantically coherent: the verbs under analysis
denote symmetrical states such as ‘be close’ or ‘be equal’, although some of them
also have dynamic uses. Conceptually, the states or actions of the symmetrical
participants are logically indistinguishable. I have demonstrated that not only does

ing
Arabic have some labile verbs, but also that there are also some peculiar cases
which conlict with the traditional notion of lability. he verbs under analysis have

lish
a causative and a non-causative use, but each of them can be either transitive or
intransitive. Traditionally, labile verbs are taken to have a transitive causative and

Pub
an intransitive non-causative use.
I argued that the reason for this peculiarity in form III is found in the semantic
component of symmetry (symmetrical action) introduced in course of the deriva-
ins
tion of form III from form I is proiled, while other components, including (non-)
causativity are not proiled (see Levin 1993; Levin, Rappaport Hovav 1993 where
am

the authors show that not only the nature of semantic components themselves, but
also their relative importance, plays a role in the syntactic behavior of the verb).
enj

his is why verbs of this pattern can be both non-causative (stative), as, for instance,
sāwā ‘be equal’, and causative, as sāwā ‘make equal, consider equal.’
As mentioned above, a derivational explanation of lability of the form III verbs
nB

under analysis is possible, but not the most reliable. It is possible to claim that the
group of verbs under analysis is derived from adjectives: as said above, deadjectival
Joh

verbs oten show lability, because both patterns are derived from non-verbal stems,
and the transitivity feature is neutralized. However Arabic deadjectival verbs con-
tain no special markers, and we cannot prove that the verbs of our group are really
-

formed from adjectives.

Arabic labile verbs are not only typologically peculiar syntactically, but also
ofs

show non-standard syntactic properties. he meaning of Arabic labile lexemes

does not have much in common with semantics of labile verbs in other languages.
pro

I am aware of no other language where symmetrical actions are a nuclear group

of labile verbs. Moreover, everyone tends to observe labile verbs which denote dy-
namic events in both uses (cf. Haspelmath 1993; Letuchiy 2006). Of course, there
ted

are exceptions: for instance, in Latin some labile verbs can denote states in their
intransitive uses: cf. variare ‘make diferent, in a diferent way; vary, be diferent.’
rec

Another conclusion I have reached is that the properties of general derivational

pattern are not relected in the properties of the concrete lexemes in any trivial
cor

way. Recall that in Arabic we deal with the property of syntactic indeiniteness
of morphosyntactic processes: most morphosyntactic derivational processes can
Un

either yield transitive or intransitive verbs. his, however, is not relected trivially
282 Alexander Letuchiy

y
in individual verbs. Otherwise we would expect to observe a large class of labile

pan
verbs in Arabic.
Let us now answer the crucial question contained in the name of the article:

Com
are Arabic form III verbs under analysis labile? he answer is yes, but in this case,
semantic and syntactic lability must be distinguished. Semantically, all verbs under
analysis are labile, because they have both a causative and a non-causative (concep-
tualized as spontaneous) use. Syntactically, they are also labile, but their variation is

ing
non-canonical: the causative use is very oten intransitive, while the non-causative
one is mostly intransitive.

lish
Pub
Acknowledgements

I am thankful to Nina Sumbatova and Vladimir Plungian, who were my advisors in diferent
periods of time, for their remarks and notes. I also want to thank the native speakers of Arabic
ins
(Ammal Hannan, Muhammad Al-Matni, Georgi Vasiliev) who were the most important par-
ticipants of this work. Finally, I have a pleasure to thank the organizers and the audience of the
am

Workshop on Verb Valency for the chance to present the work in very friendly and working
atmosphere.
enj
nB

References

Benmamoun, Elabbas, Aoun, Joseph E. & Choueiri, Lina. 2010. he Syntax of Arabic. Cambridge:
Joh

CUP.
Blachère, Régis & Maurice Gaudefroy-Demombynes. 1952. Grammaire de l’arabe classique. Paris:
G. P. Maisonneuve.
-

Blachère, Régis. 1958. Elements de l’arabe classique. Paris: G. P. Maisonneuve.

Bybee, Joan. 1995. Regular morphology and the lexicon. Language and Cognitive Process 10:
425–455. doi: 10.1080/01690969508407111
ofs

Comrie, Bernard. 2006. Transitivity pairs, markedness, and diachronic stability. Linguistics 44(2):
303–318. doi: 10.1515/LING.2006.011
Creissels, Denis. 2014. P-lability and radical P-alignment. Linguistics 52(4): 911–944.
pro

doi: 10.1515/ling-2014-0012
Daniel, Michael A., Ganenkov, Dmitri V. & Merdanova, Solmaz R. In press. Causatives in Agul.
doi: 10.1075/slcs.126.04dan
ted

Dixon, Robert M. W. 1980. Ergativity. Cambridge: CUP.

Fassi Fehri, Abdelkader. 2012. Key Features and Parameters in Arabic Grammar [Linguistik
rec

Aktuell/Linguistics Today 182]. Amsterdam: John Benjamins. doi: 10.1075/la.182

Galiamina, Yulia. 2006. Transitivity and lability in Songhay. In Case, Valency and Transitivity
[Studies in Language Companion Series 77], Leonid I. Kulikov, Andrei L. Malchukov & Peter
cor

de Swart (eds), 359–374. Amsterdam: John Benjamins. doi: 10.1075/slcs.77.22gal

Grande, Bentsion M. 1998. Kurs arabskoj grammatiki v sravnitel’no-istoričeskom osveščenii (Course
Un

of Arabic grammar in historical and comparative perspective). Moscow: Russkie slovari.

 Arabic ‘labile verbs’ in form III 283

y
Haspelmath, Martin. 1993. More on the typology of inchoative/causative verb alternations. In

pan
Causatives and Transitivity [Studies in Language Companion Series 23], Bernard Comrie &
Maria Polinsky (eds), 87–120. Amsterdam: John Benjamins. doi: 10.1075/slcs.23.05has
Hopper, Paul & hompson, Sandra. 1980. Transitivity in Grammar and Discourse. Language

Com
56(2): 251–299. doi: 10.1353/lan.1980.0017
Jakovlev, Nikolaj F. & Ašhamaf, Daud A. 1941. Grammatika adygejskogo jazyka (A grammar of
Adyghe). Moscow: Nauka.
Kehayov, Petar. In press. Labile verbs in Estonian.

ing
Kibrik, Alexander E., Muravyeva, Irina A. & Kodzasov, Sandro V. 2001. Jazyk i fol’klor alutorcev
(Alutor language and folklore). Moscow: Jazyki slavjanskoj kul’tury.
Khrakovskij, Viktor S. 2011. Kauzativnaja derivacija: osobyj slučaj (he causative derivation:

lish
A special case). In Meždunarodnaja konferencija, posvjaščennaja 50-letiju Peterburgskoj
tipologičeskoj školy: Materialy i tezisy dokladov (International conference, dedicated to the

Pub
50th anniversary of Saint-Petersburg Typological School. Proceedings and abstracts), Viktor
S. Khrakovskij et al. (eds), 184–192. Saint-Petersburg: ILI RAN.
Kulikov, Leonid I. 1999. May he prosper in ofspring and wealth: A few jubilee remarks on the
typology of labile verbs and Sanskrit púsyati ‘prospers; makes prosper’. In Tipologija i teo-
ins
rija jazyka: Ot opisanija k objasneniju (Typology and language theory: From description to
explanation), Yakov G. Testelets & Ekaterina V. Rakhilina (eds), 224–244. Moscow: Jazyki
am

russkoj kul’tury.
Kulikov, Leonid I. & Lavidas, Nikolaos (eds). 2014. Typology of Labile Verbs: Focus on Diachrony.
Special issue of Linguistics 52(4).
enj

Letuchiy, Alexander B. 2006. Tipologija labil’nyx glagolov: semantičeskije i morfosintaksičeskije

aspekty (Typology of labile verbs: Semantic and morphosyntactic aspects). PhD dissertation.
nB

Moscow.
Letuchiy, Alexander B. 2010. Lability and spontaneity. In Transitivity: Form, Meaning, Acquisition,
and Processing [Linguistik Aktuell/Linguistics Today 166], Patrick Brandt & Marco García
Joh

García (eds), 237–256. Amsterdam: John Benjamins. doi: 10.1075/la.166.10let

Letuchiy, Alexander B. 2013. Tipologija labil’nyx glagolov (Typology of labile verbs). Moscow:
Yazyki slavjanskoj kul’tury.
-

Levin, Beth & Rappaport Hovav, Malka. 1995. Unaccusativity. At the Syntax-lexical Semantics
Interface. [Linguistic Inquiry Monographs 26]. Cambridge MA: he MIT Press.
Levin, Beth. 1993. English Verb Classes and Alternations. Chicago IL: Chicago University Press.
ofs

Malchukov, Andrej L., Haspelmath, Martin & Comrie, Bernard (eds). 2010. Studies in Ditransitive
Constructions: A Comparative Handbook. Berlin: Mouton de Gruyter.
Nedjalkov, Vladimir P. 1969. Nekotoryje verojatnosnyje universalii v glagol’nom slovoobrazovanii
pro

(Some probabilistic universals in verbal derivation). In Jazykovyje universalii i lingvističes-

kaja tipologija (Language universals and linguistic typology), Igor F. Vardul’ (ed.), 106–114.
Moscow: Nauka.
ted

Næss, Åshild. 2007. Prototypical Transitivity [Typological Studies in Language 72]. Amsterdam:
John Benjamins. doi: 10.1075/tsl.72
rec

Nichols, Johanna, Peterson, David A. & Barnes, Johnatan. 2004. Transitivizing and detransitiv-
izing languages. Linguistic Typology 8: 149–211. doi: 10.1515/lity.2004.005
Prost, Alain. 1956. La langue songay et ses dialectes. Dakar: IFAN.
cor

Rožanskij, Fedor I. 1991. O tranzitivnosti i intranzitivnosti v jazyke songaj (On transitivity and
intransitivity in Songhay). In Afrika: Kul’tura, etničnost’, jazyk, 179–187. Moscow: Institut
Un

jazykoznanija.
284 Alexander Letuchiy

y
Saad, George Nehweh. 1982. Transitivity, Causation and Passivization. London: Kegan Paul

pan
International.
Scheka, Yury V. 1999. Intensivnyj kurs tureckogo jazyka (An intensive course of Turkish). Moscow:
Izdatel’stvo MGU.

Com
Tsunoda, Tasaku. 1985. Remarks on transitivity. Journal of Linguistics 21: 385–396.
doi: 10.1017/S0022226700010318

ing
Author queries

lish
Please provide a complete reference for the citation (Daniel, Michael A., Ganenkov,
Dmitri V. & Merdanova, Solmaz R.) in References of this chapter.

Pub
Please provide a complete reference for the citation (Galiamina, Yulia. 2006.) in
References of this chapter.
Please provide a complete reference for the citation (Jakovlev, Nikolaj F. & Ašhamaf,
ins
Daud A. 1941.) in References of this chapter.
Please provide a complete reference for the citation (Kehayov, Petar.) in References
am

of this chapter.
Please provide a complete reference for the citation (Khrakovskij, Viktor S. 2011.)
enj

in References of this chapter.

Please provide a complete reference for the citation (Levin, Beth & Rappaport
nB

Hovav, Malka. 1995.) in References of this chapter.

Please provide a complete reference for the citation (Malchukov, Andrej L.,
Haspelmath, Martin & Comrie, Bernard (eds). 2010.) in References of this chapter.
Joh
-
ofs
pro
ted
rec
cor
Un

View publication stats