Cleft constructions and focus strategies in Modern Armenian

In: Faits de Langues
View More View Less
Full Access


Cleft constructions are one of the possible focus strategies available in Modern Armenian alongside prosody and specific syntactic constructions. Cleft constructions in Modern Armenian are biclausal constructions with a matrix clause and a relative-like clause, with an identificational clause as a matrix clause headed by a copula (in present or past), while in the relative-like clause introduced by the main subordinator, the relativized argument is coindexed with the argument of the copula. Though typologically cleft constructions are considered typical of languages with rigid word order, they are common in Modern Armenian, a language with flexible word order. It is argued that the intensity of focalization depends on the strategy used, with simple prosody marking associated with the lowest level of intensity, and preverbal position and clefts associated with intermediate and high-intensity focalization respectively. The corpus-based data show an unequal distribution of clefted pronouns as predicate clefts (impersonal with no agreement) and subject clefts (copular verb coindexed with personal pronouns as a subject) depending on the person and the polarity. The existence of cleft-like constructions in Classical Armenian and both Modern Armenian standards is argued to be evidence of diachronic continuity and a possible grammaticalization path from cleft constructions to the auxiliary movement focus strategy.


Cleft constructions are one of the possible focus strategies available in Modern Armenian alongside prosody and specific syntactic constructions. Cleft constructions in Modern Armenian are biclausal constructions with a matrix clause and a relative-like clause, with an identificational clause as a matrix clause headed by a copula (in present or past), while in the relative-like clause introduced by the main subordinator, the relativized argument is coindexed with the argument of the copula. Though typologically cleft constructions are considered typical of languages with rigid word order, they are common in Modern Armenian, a language with flexible word order. It is argued that the intensity of focalization depends on the strategy used, with simple prosody marking associated with the lowest level of intensity, and preverbal position and clefts associated with intermediate and high-intensity focalization respectively. The corpus-based data show an unequal distribution of clefted pronouns as predicate clefts (impersonal with no agreement) and subject clefts (copular verb coindexed with personal pronouns as a subject) depending on the person and the polarity. The existence of cleft-like constructions in Classical Armenian and both Modern Armenian standards is argued to be evidence of diachronic continuity and a possible grammaticalization path from cleft constructions to the auxiliary movement focus strategy.

1. Introduction1

Typologically, cleft constructions are considered to be characteristic of languages with rigid word order (Jespersen 1927; Lambrecht 2001]. However, in Modern Armenian (hereafter Armenian unless indicated otherwise), a pro-drop language with flexible word order, such constructions exist and can be quite frequent, as in other typologically comparable languages of the region (cf. Kazenin 2001; Erschler 2012; Komen 2015; Forker 2021; as well as others in this volume).

Several papers exist on the preverbal focus strategy in Eastern Armenian (cf. Comrie 1984; Tamrazian 1991; Megerdoomian 2011; Kahnemuyipour & Megerdoomian 2017; Semenova 2014a, 2014b). However, to our knowledge, no research has been done on cleft constructions in Armenian, and Western Armenian data have never been analyzed from the perspective of information structure. Moreover, Dum Tragut (2009: 625) claims overtly that: “Clefting as a type of left dislocation is not exhibited” in Eastern Armenian.

In this paper, an overview of cleft and cleft-like constructions in Armenian is presented, as well as other major information prominence marking strategies such as prosodic means and the verb/auxiliary movement strategy. The main types of cleft constructions (CC) with their characteristic features are outlined, and the distinctive discursive and functional properties of clefts as compared to other information prominence marking strategies are presented from an intravariational (two Modern Armenian standards) and typological perspectives. It is argued that clefts are diachronically prior to the verb/auxiliary movement strategy and that the latter could be a result of the grammaticalization of CCs reinforced by areal effects. Corpus-based data and statistics were used to back up the research results as often as possible.

Modern Armenian has two standards: Modern Eastern Armenian (MEA in examples), the language spoken in today’s Republic of Armenia and the Armenian communities of Iran and the ex-Soviet republics, and Modern Western Armenian (MWA in examples), which is mainly spoken by the Armenian communities in the Middle East, as well as by the traditional Armenian communities in Europe, the USA and Latin America. Classical Armenian is preserved for canonical uses. Besides the two standards, a number of Armenian dialects complete the Modern Armenian language continuum.

Areally, Eastern Armenian has been mainly in contact with Iranian, Caucasian and some Turkic languages, as well as with Russian in more recent times, whereas Western Armenian has been in close contact with Turkish and Kurdish languages and to some extent with Levantine Arabic, occupying a buffer position between the Caucasian, Mesopotamian and Balkan Sprachbunds (cf. Haig 2017; Friedman 1996; Donabedian and Sitaridou 2020). More recent contacts for Western Armenian are some European languages (e.g. French, English etc.).

The history of the Armenian language displays several typological shifts, from Classical Armenian (since the 5th century AD) to both modern standards, by way of Middle Armenian, which is well documented for Western Armenian, due to socio-political conditions, but traced much more obscurely for Eastern Armenian.2

Morphologically, Classical Armenian was predominantly a flexional language, with mainly synthetic verbal forms and a very few periphrastic constructions with the auxiliary verb ē “be”, that is, only the indicative perfect and plus-quam-perfect were periphrastic constructions composed of the perfect participle and the auxiliary verb ē “be’. With some slight changes, these two tenses are preserved in both modern standards as periphrastic constructions.

As the old indicative present and imperfect forms shifted to the subjunctive (present and past respectively) in both Eastern and Western Armenian, the two modern standards innovated their verbal systems in different ways. Western Armenian created new indicative tenses by means of the adjunction of particles (mainly proclitics) to the forms, keeping the perfect and plus-quam-perfect as the only indicative periphrastic verbal forms using an auxiliary, whereas Eastern Armenian adopted auxiliary-based periphrastic forms for almost the whole indicative system (see Donabedian & Ouzounian 2008; Donabedian 2018a: 116–117; for more details on TAM forms in Eastern Armenian see Bybee 1994, Plungian 2018).3 This generalization of periphrastic forms with the auxiliary in Eastern Armenian is the ground which made possible the generalization of a specific feature involved in auxiliary/copula movement focalization strategy in this standard.

Typologically, Modern Armenian is a left-branching language with nominative-accusative alignment, a mostly agglutinative nominal system, and a more fusional verbal system. It has flexible word order with some features typical of OV languages.

Unless indicated otherwise, Eastern Armenian data come from the Eastern Armenian National Corpus (, a comprehensive corpus of about 110 mln. tokens), those of Western Armenian come from the Digital Library of Armenian Literature ( and a Nooj-based Western Armenian corpus (Anaid Donabedian, unpublished). Classical Armenian data are taken from the online resource of the Classical Armenian Bible with Parallel King James Version (

The paper is composed of an introduction, two sections and a conclusion. Section 2 covers focus marking strategies in Armenian, with 2.1 on the expression of focus in Armenian and 2.2 on constituent order and the preverbal focus position. Section 3 discusses clefts and cleft-like constructions, with 3.1 on distinctive properties of clefts in Armenian, 3.2 on syntactic roles of clefts, 3.3 on predicate argument vs subject argument in clefts, 3.4 covering cleft-like constructions, 3.5 presenting data on cleft constructions in Classical Armenian, and finally, 3.6 discussing clefts vs other focus marking strategies.

2. Focus marking strategies in Armenian

2.1. The expression of focus in Armenian

In this paper we adopt the definition of focus proposed by Lambrecht:

… the focus of a sentence, or, more precisely, the focus of a proposition expressed by a sentence in a given utterance context, is seen as the element of information whereby the presupposition and the assertion differ from each other. The focus is that portion of a proposition which cannot be taken for granted at the time of speech. It is the unpredictable or pragmatically non-recoverable element in an utterance. The focus is what makes an utterance into an assertion”. (Lambrecht 1994: 207)

While analyzing the information structure in Armenian, two basic types of foci can be distinguished: 1. Default/nuclear focus, which often corresponds to the prosodically most prominent constituent, identifiable by the nuclear stress in any sentence even with unmarked information structure. 2. Marked/exhaustive focus, with a higher degree of prosodic prominence. This kind of focus may have different context-dependent meanings (contrast or identification being the most frequent one) and implies specific marking strategies. Both the nuclear and marked foci in our examples are marked in uppercase, whereas elements associated with the issues under discussion are marked in bold.

In Armenian, three main marked focus strategies can be outlined, with more or less relevance for Eastern and Western Armenian:

  1. Prosody focus (pitch accent),
  2. Constituent order (preverbal syntactic focus position),
  3. Cleft constructions.

Some of these strategies, especially prosody marking, can overlap with others. In-situ prosodic focus marking is represented by prosodic stress on the focused constituent without any word order changes.

There are also certain focus-sensitive particles with an additional semantic load (e.g. hench “just”, isk “exactly”, al “also” (MWA), ēl “also” (MEA) etc.). They can be prepositive or postpositive, and may or may not be cliticised. Of the focus-sensitive particles in Armenian, the particle al/ēl “also” is by far the most frequent, especially in oral discourse. This particle is always encliticised, and requires auxiliary/copula movement in Eastern Armenian if an analytic tense form is used. The auxiliary/copula in its turn is an enclitic and the chain of two enclitics has always fixed order, with the particle coming first and the auxiliary/copula following it (1), (2).4


2.2. Constituent order and the preverbal focus position

Modern Armenian is a pro-drop language with relatively flexible word order, and both SOV and SVO have been argued to be the basic word order in Armenian, depending on analyses. Western Armenian is a predominantly verb-final SOV language (Donabedian 2010; 2018b), whereas in Eastern Armenian, the distribution of SVO and SOV is more proportional (Samvelian et al. submitted), with SVO order considered more distant and formal (e.g. more frequent in press (3), in non-fiction discourse or formal discourse etc.), and allowing post-verbal focus.


In Eastern Armenian, SOV is especially frequent in verbal phrases with a bare object (4). Such phrases can be viewed as one semantic unit, and share some properties with complex predicates, which are quite characteristic of Eastern Armenian, especially under the influence of Persian (e.g. lach linel “cry” (lit. “cry to be”), xaġ anel “play” (lit. “play to do”), par gal “dance” (lit. “dance to come”), etc.).6


Whenever the verb is used in a non-finite form, verbal arguments are usually preverbal (5a), and VO order is often ungrammatical (5b), which is further evidence for head final order in MA.


The question of basic word order and the possibility of having more than one basic word order is also of typological and areal relevance (see Skopeteas & Fancelow [2010] for Georgian, in which both SVO and SOV are possible), and worthy of a comprehensive corpus-based study.

Preverbal position is considered to be the default focus position in Armenian (6a-7), based on mainly Eastern Armenian data (cf. Comrie 1984; Tamrazian 1991; Megerdoomian 2011; Kahnemuyipour & Megerdoomian 2017; Semenova 2014a, 2014b). Ongoing exploratory study (Donabedian 2018b) puts forward the same thesis for Western Armenian, which is consistent with areal (cf. for Turkish Göksel and Özso 2000) and typological evidence (cf. Sornicola (2006: 380) among many others for the correlation of SOV and preverbal focus position).


The neutral word order of periphrastic constructions in both standards is converb plus auxiliary verb for affirmative forms, and negative auxiliary and converb for negative forms (6b-8).7 In both Armenian standards, it is the verb ē “be” that functions as both an auxiliary and a copula verb with its present and past tenses. In the case of periphrastic constructions (as well as in copular predications), the auxiliary verb is never stressed, except when it is used in its negative form. In other words, the auxiliary/copula is an enclitic whenever used in its affirmative form, and follows the default nuclear stress position, but attracts the nuclear stress when negative.8


Examples (6b) and (8) are compatible with a neutral information structure context: the negative auxiliary bears the nuclear stress, assuming “default focus” as defined above. In MEA, if there is a more prominent constituent in the utterance, as in (6c), in which es “I” is a contrastive focus, the negative auxiliary loses its stress together with the nuclear stress. In (6d) it is the subject es “I” that is stressed, and the negative auxiliary čhem functions as a clitic similar to the affirmative auxiliary forms. Instrumental phonetic research would be necessary to explore this phenomenon further.


Important syntactic differences exist between Eastern and Western Armenian as regards the properties of periphrastic constructions, which are more prevalent in Eastern than in Western Armenian. In Eastern Armenian, with periphrastic verbal forms, it is no longer the lexical verb, but the in-situ focus which hosts the clitic auxiliary. In other words, in-situ focus leads to a movement of the auxiliary from its default post-converbial position to a post-focus position (6d, 6e).9


The informational character of the focus seems sensitive to the syntactic role of the preverbal element. Core arguments can both precede and follow the verb, with obligatory auxiliary movement when they are focused in pre-converbial position. However, preverbal position does not necessarily indicate contrastive focus (9b), e.g. in the case of bare objects or destination (9a, 9b, 9c) etc.10


Both in Eastern and in Western Armenian, when the verb ē “be” is used as a copula, it is placed after the nominal and adjectival predicate (10a), whether negative or affirmative. The simple inversion of the copula is used for focus marking with affirmative and negative forms of the copula with clitic prosody in Eastern Armenian (10b). Unlike in Eastern Armenian, in Western Armenian the preverbal focus marking strategy is possible only with nominalization of the adjectival predicate (10c).


In contrast to the auxiliary use of the verb ē “be”, it is ungrammatical to have an inverted stressed negative copula (10d).


As has already been mentioned, bare objects in Armenian tend to be placed before the verb, as is the case for complex predicates (11). Similar to complex predicates, in MEA, verb phrases with a bare object have the auxiliary verb between the bare object (or the nominal constituent in case of complex predicates) and the converb when used in a periphrastic tense (12).11


Another case of obligatory auxiliary movement concerns questions with interrogative words (what, who, where etc.) (13a). It is ungrammatical to keep the auxiliary after the converb (13b).12


In contrast to Eastern Armenian, auxiliary verb movement in Western Armenian is ungrammatical in periphrastic constructions (cf. (14) and (15) with an interrogative word and a preverbal bare DO, which are in focus by default and there is no auxiliary verb movement).


Auxiliary verb movement in Eastern Armenian can be considered an areal feature (cf. certain Caucasian languages [Udi, Tsakhur, Lak, Itsari, Bagvalal) and Iranian languages [Tat, Talyshi, Gilaki]), in which such a strategy exists but does not completely coincide with that of Eastern Armenian (cf. Stilo 2008; Megerdoomian 2011; Semenova 2014b; Forker 2021). As mentioned in the introduction, Western Armenian is more influenced by Turkish being part of the Balkan Sprachbund, whereas Eastern Armenian is part of the Caucasus-Iran linguistic area. This areal distribution could account for the lack of the auxiliary movement strategy in Western Armenian.

As suggested above, it is also possible that TAM formation differences between Western and Eastern Armenian, especially in the indicative mood, resulted in some preferences in focus strategies in each standard. As we have seen, Eastern Armenian is much more innovative in this domain, with a new system of periphrastic forms, whereas Western Armenian is more conservative, as even the new TAM forms are often based on ancient Classical Armenian ones (e.g. the present indicative (Donabedian 2018a). The large array of uses of the auxiliary verb in Eastern Armenian could in itself be a stimulus for the development of more sophisticated grammaticalized strategies. The areal factor could incite but not necessarily cause them. In the following chapters it is argued that the auxiliary movement strategy in Eastern Armenian could be the result of a grammaticalization path of cleft constructions.

3. Clefts and cleft-like constructions

3.1. Distinctive properties of cleft constructions

Cleft constructions are usually understood as:

…a complex sentence structure consisting of a matrix clause headed by a copula and a relative or relative-like clause whose relativized argument is coindexed with the predicative argument of the copula. Taken together, the matrix and the relative express a logically simple proposition, which can also be expressed in the form of a single clause without a change in truth conditions. (Lambrecht 2001: 467).

Initially, CCs were analyzed on the basis of examples from English and French, which have very consistent syntactic properties for such constructions. One of the main challenges of cleft analysis from a typological perspective is, therefore, defining the proper set of semantic and syntactic properties delimitating the target constructions.

Clefts in Armenian are biclausal constructions with a matrix clause and a relative-like clause. The matrix is an identificational clause headed by a copula (in present or past) and in the relative-like clause, the relativized argument is coindexed with the argument of the copula as in (16, 17). As in the definition of Lambrecht (2001: 467) above, in Armenian, the main propositional content of clefts, as expressed in the relative-like clause, could also be expressed as a single clause with no change in truth conditions.


In Armenian, there are other constructions that share some features with clefts (e.g. relative clauses, purpose clauses, condition clauses etc.); however, the following criteria can be distinctive for clefts:

  1. Agreement. In cleft constructions the relativized argument is coindexed with the argument of the main clause, which may be either a predicative argument (any constituent) or a subject argument (with personal pronouns only) of the copula. In the case of predicative argument constructions, which are equivalent to those with empty subjects observed in English and French, the copula does not agree with the argument. This feature is not displayed uniformly, as will be shown further on.
  2. Prosody. In cleft constructions there is a mandatory identificational or contrastive focus accent on the clefted constituent, while in regular relative constructions the sentence stress is positioned freely according to the information structure.
  3. Pragmatics. The simple deletion of the copula of a CC’s matrix clause together with the relativizer does not influence the logical proposition in its truth conditions. In the case of a relative clause or other similar construction, such manipulation cannot result in a proposition with the same truth conditions (cf. the relative clause in (18a, 18b) with a cleft and its simple proposition (18c, 18d) as well as (16, 17)).


  1. Complementizer. While the relative clause can be introduced either by a set of flectional relative/interrogative pronouns, or by the invariable relativizer/complementizer or (pronounced as [vor]) “that” (for more details on relative clauses in Modern Armenian see Hodgson 2019, among others), for clefts, only the latter is possible. Despite its shared origins with the interrogative pronoun/adjective or “which” and used as an inflected relative pronoun, invariable or is a different marker, serving as a generic complementizer in Armenian with a very wide range of syntactic/semantic uses: relativizer in a relative clause; conjunction /complementizer (factive/ non-factive, purpose, time, condition, reason); modal /discourse marker (skepticism, doubt, astonishment, focus etc.) etc.13

In cleft constructions, the matrix clause with the clefted constituent is always placed before the relative-like clause, and it is often placed at the beginning of the sentence (17,18c), though clefts can also split the sentence to have any nominal element clefted (16). In Armenian, the clefted constituent is adjacent to the enclitic copula followed by the complementizer introducing the relative-like clause. No split is possible between these three constituents. The pre-copular position of the clefted constituent is in line with the preverbal focus marking strategy according to which the focused constituent is always in a preverbal position.

As mentioned above, the auxiliary/copula always functions as a clitic when affirmative, and can be stressed only when negative. The negative auxiliary/copula is cliticised only after a focused constituent. Unlike in Eastern Armenian, negative auxiliaries/copulas in Western Armenian are always stressed, even when following the focused constituent element.

3.2. Syntactic roles of clefts in Armenian

In Armenian there is no particular restriction concerning the syntactic role of the clefted constituent, apart from the verb/predicate, as in other languages (cf. Nichols 1994: 76). Examples (19, 20: S) and (21, 22: DO) show some core arguments clefted in Western and Eastern Armenian respectively.


Adverbial modifiers are one of the categories that are frequently clefted in Armenian (23, 24, 25).


Clefts with an adverb/ adverbial phrase as a clefted constituent are one of the most frequent types not only in Armenian (26, 27), but also typologically. This type of clefts mainly involves clefted manner, place or time adverbials. Overall, 7570 matches have been found for ‘ADV + COP + or’, with about 40% involving present affirmative and negative copulas, 15% with a past copula, and only 4 % with a past negative copula. This is quite a significant number, given that there should not be many irrelevant examples. According to the results of EANC, in certain cases a clear preference can be discerned as regards the form of the copula, e.g. there is a clear preference for the affirmative copula in most time adverbials (e.g. the search query “quantifier + tari “year” + COP + or “that”, meaning “it’s been X years since…; for X years now…” has 642 matches with the affirmative copula vs. 39 with the negative one). On the contrary, the negative copula prevails with the adverb mišt “always” (1156 matches with negative copula vs 38 matches with the affirmative one). This can be accounted for by the inherent properties of mišt “always” as a modifier of the truth conditions of the verb rather than an identifier of a circumstance.14


Time clefts are formed like nominal clefts, and yet semantically and formally they are different. Two main types of time CCs can be distinguished:

  1. Time indication clefts (28), which highlight a particular time point or time period, e.g. now, yesterday, last year, next week etc.


  1. The second type involves durative clefts (29a), which designate a certain span of time. These constructions correspond to French “cela fait Xt que…”. The difference in Armenian as compared to French is that the non-focused option of the same proposition can be expressed simply by omission of the copula and the complementizer (29b) (cf. French, where the simpler proposition is expressed by other constructions, e.g. “depuis Xt).


Unlike in the main types of clefts, in this particular kind, the complementizer or “that” is not obligatory, thus the relative-like clause can be introduced by the conjunction inčh “what, that” (30a) (possible only in MEA), zero marking (30c), or by the complementizer or “that” (30b). Thus, even taking into account the possibility of (29b) above, which could mean that, unlike the similar construction in French, the Armenian construction is a genuine cleft, it seems that in Armenian, this construction is also to some extent idiomatic.


Another particularity of durative clefts is the possibility of having the verb linel “be” (31) in all tense forms or the defective verb ka “exist” as a copula (32a), as opposed to the exclusive use of the defective verb ē “be” in all other clefts (cf. the verb linel “be” as a copula in the aorist (32b)).15 Both linel “be” and ka “exist” have more stative semantics than ē “be”.


3.3 Predicate argument vs subject argument in clefts

Unlike in English or French, the matrix clause of Armenian clefts does not contain an empty pronoun (e.g. it [is]… (Eng.), c’[est]... (Fr.) etc.). The focused constituent functions as a predicate, and it can have an impersonal construction reading. Agreement of the copula with the clefted constituent and the respective tense in the relative-like clause is not mandatory. It is possible to have complete agreement, as in (33, 34), where the copula agrees in number and tense (see the remark on the status of tense variation at the end of this section).


However, the copula with default third person singular present is also possible (cf. (35) in which the clefted constituent is in plural with a copula in singular and the predicate in the relative-like clause is in the aorist).


In the same way, cleft personal pronouns can have the default third person copula (19, 36, 37, 38).


The default 3rd person singular copula is in line with the default impersonal copular form, as well as the default present rather than sequenced past form, which can be accounted for by the lack of sequence of tenses in the strict sense in Armenian (39).


However, it is also possible to have a syntactic subject of the matrix clause with the copula inflected (40, 41, 42).


Table 1 and Table 2 focus on the distribution of clefts with a personal pronoun clefted as a predicate and as a subject respectively.


The results show that negative copula forms are inversely proportional to the general tendency of a clefted subject rather than a clefted predicative. Namely, the matches for the first-person singular es “I” and menkh “we” are 86% and 83% respectively, and 67% and 65% for du “you” in singular and plural respectively. Such results could be accounted for the ‘assimilation/attraction’ phenomenon, i.e. when the personal pronoun and the copula are adjacent, the speakers tend to use the appropriate personal conjugated copula form, whereas in case of negation, the personal pronouns being ‘separated’ from the copula by the negative prefix, the attraction is less significant. As the auxiliary/copula is a clitic in MEA, the sequence of a personal pronoun plus the appropriate agreeing auxiliary/copula form tends to be more available to speakers than that with a 3rd person singular auxiliary. Besides, forms where the second person singular and plural appear as clefted subjects are about three times more frequent than those in which they appear as clefted predicates (84% vs 16 %, and 79 % vs 21 % respectively). Another interesting result is the almost total absence of the past tense copula (both affirmative and negative) with personal pronouns clefted as predicates (Table 1).

Number and person marking in constructions with clefted personal pronouns reflect the agreement of the copula with the subject of the main clause. The examples also display tense variation, which is not to be ascribed to agreement, but rather depends on discursive features: In cases where we have a past relative-like clause, time co-indexation of the copula in the main clause marks the anchoring of the whole cleft sentence in the historical narration, while the lack of co-indexation, as in examples (19) and (20) (present main clause and past relative-like), marks the anchoring of the main clause in the time of enunciation. This could be demonstrated with wide-context examples, but the limits of the present paper do not allow us to do it here.

3.4. Cleft-like constructions in Armenian

It is challenging to question the boundaries of the domain of clefts, especially with respect to a concurrent prenominal, non-finite strategy for relativization (43a, 43b), well known in the typology of relative clauses and well-represented in the Caucasus-Anatolia continuum in general (cf. Gandon 2016) and in Armenian, in particular (cf. Hodgson 2019). This strategy is available as an alternative to the basic cleft constructions for the focalization of any personal pronoun or core argument NP (S or, to a lesser extent, DO) and it is very frequent in Armenian, with different possible orders (43a, 43b).


An important set of properties distinguish the non-finite constructions from regular cleft constructions. Unlike basic clefts, the non-finite strategy is monoclausal, with syntactic restrictions (equivalent to subject clefts, and possibly some object clefts), and it has a relatively free word order. The word order difference between canonical CCs and their non-finite equivalents results in different prosodic contouring of these two constructions. While in finite clefts the relative-like clause always follows the sentence stress (borne by the initial focalized constituent) and receives the prosodic contour of backgrounded information, in the monoclausal non-finite constructions, by default the focalized constituent is in the pre-final position (43a), and the construction marks a lower degree of focalization than a biclausal finite cleft, although the alternative order is possible too (43b). Therefore, despite the functional similarity with canonical cleft constructions, the non-finite constructions would be more appropriately classified as cleft-like constructions.

Besides non-finite relative clauses, a range of constructions (44, 45, 46, 47) have similarities with clefts in Armenian.


One of these constructions introduced by or contains a matrix clause, often with a copula (in the negative or in a rhetorical question) and a subordinate clause, and semantically seems close to a clause of purpose (cf. (48) in which the complementizer or can have a purpose reading, i.e. “so that, in order to”).


These constructions are often introduced by the emphatic discursive marker ba “and, so” as a rhetoric question (49, 50) (over 500 matches for such constructions introduced by ba are found in EANC). When the matrix clause has a negative equative verb, the subordinate clause predicate is often in the subjunctive. Traditionally, the subordinate clauses in such constructions are considered to be attributive (Abrahamyan 2004: 155).


In Eastern Armenian oral discourse, it is quite frequent to have questions with a final or (51, 52). Most probably the or in such interrogative phrases is the same that in the constructions mentioned above expressing some implied reason/argumentation, i.e. in such constructions, the proposition expressed by the matrix clause is pragmatically the inverse.


Another type of construction is composed of a matrix clause containing a subject (often expressed by a pronoun) with an equative verb ē “be” and a subordinate clause introduced by or with only a stative equative verb kam “be” coindexed with the matrix clause (53). The construction has an emphatic identifying semantics, which is often intensified by the adverb hench “exactly, just”.


Interestingly, a similar construction with two mirrored synonymous existential verbs exists in Azeri (54), but not in Western Armenian, Persian or Turkish.16 This could be strong evidence for areal contact effects; however, more areal data are necessary to make such conclusions.


All the constructions presented in this chapter are considered as cleft-like constructions, since they deviate in a way or another do not have all the distinctive features of clefts in Armenian given above. More detailed study is needed to enable to determine whether or not each of these constructions should be considered full-fledged cleft constructions or not.

3.5. Clefts in Classical Armenian

Diachrony provides interesting evidences for a holistic analysis of clefts and their interaction with other focus-marking strategies in MA. Similar constructions (55a, 55b) are found in Classical Armenian.


In example (55a), the focused constituent is es “I”, which is preceded by negation and followed by inčh “thing, what”. The status of the latter inčh “thing, what” is often ambiguous in Classical Armenian texts, and it is considered to be a simple Greek calque or an emphatic marker (56a, 56b).


Despite the ambivalent character of inčh in certain contexts in Classical Armenian, the parallels with cleft constructions and some cases of inčh seem convincing. This can be reinforced by English, French and Eastern Armenian translation equivalents, which use cleft constructions. Besides, in Armenian, inčh can function as a correlative pronoun/conjunction, and it can even replace the relativizer or in cleft-like constructions of time in Eastern Armenian (30).

Clefts are considered to be easily borrowed (one of the hypotheses concerning the origin of clefts in North East Caucasian languages is contact with the languages of the region, namely, Armenian among others, Harris [2001: 161]). However, Classical Armenian data show the continuity of CCs in Armenian and mark their inherent rather than borrowed character in Armenian. Besides, the presence of clefts in Classical Armenian could also be evidence for a grammaticalization path of clefts into Modern Armenian. According to Harris and Campbell (1995: 166) typologically “Monoclausal highlighting constructions often originate as biclausal structures – clefts or anti-clefts”. Harris and Campbell (1995) propose three stages for such a development path, with the first being a biclausal structure, which then develops into a mixed type of structure with some characteristics of a biclausal and some of a monoclausal one, and finally becomes a monoclausal structure. The development of the auxiliary/verbal movement strategy in Eastern Armenian could be the result of a similar grammaticalization path. Such an evolution could possibly be reinforced by areal contact effects. This becomes even more visible in the context of differences between Western and Eastern Armenian, with Western Armenian having preserved only biclausal and participial cleft constructions, whereas Eastern Armenian has developed the monoclausal auxiliary /copula movement strategy parallel to biclausal and participial cleft constructions.17

3.6. Clefts vs other focus marking strategies in Armenian

As shown above, in Eastern Armenian clefts are one of three available focus strategies: 1. Prosody marking (57a); 2. constituent order marking (preverbal syntactic focus position) (57b) and 3. cleft constructions (57c) (cf. “postverbal focus” in Georgian (Skopeteas and Fancelow 2010).18


It could seem that, typologically, having clefts alongside other syntactic focus marking strategies is somewhat unexpected. However, Armenian is not the only language, at least in the region concerned, to exhibit various possibilities (cf. Kazenin 2001; Erschler 2012, Komen 2015; Forker 2021; and others in this volume).

One possible explanation for the existence of clefts in Armenian despite the variety of other available focus marking strategies is that clefts mark the focus syntactically and unambiguously, even when prosody is not available. Thus, this is a preferable focus marking strategy in written discourse as compared to the in-situ focus marking strategy, which can have different readings as long as the prosody is not present. As has been shown, the preverbal strategy can also mark a constituent unambiguously, as long as there is auxiliary/copula movement. This strategy being relevant only for Eastern Armenian, clefts would be preferred especially in Western Armenian, as well as in Eastern Armenian when synthetic verb forms are used. The EANC data show a clear preference for clefts in Eastern Armenian written vs oral discourse (70% vs 30% respectively). More data would be necessary to check the hypothesis of discourse-conditioned CC distribution in Eastern and Western Armenian.

4. Conclusion

Clefts in Armenian are grammaticalized constructions with a high frequency of use, which allow the clefting of any constituent of the sentence except the verb. The present exploratory study highlights some issues that are of interest from the perspective of Armenian as well as that of the typology of clefts.

Clefts are one of a number of possible focus strategies available in Armenian involving prosody and specific syntactic constructions based on word order or subordination mechanisms. The question of basic word order in Armenian (SOV or SVO) is disputed with clear evidences about a predominantly head final order. Certain orders can be identified with regard to information structure, namely, the preverbal (or pre-final) and postverbal positions for different types of focus marking. Despite some minor differences, both Eastern and Western Armenian display an asymmetry between affirmative and negative periphrastic constructions. The generalization of periphrastic constructions (more radical in Eastern Armenian than in Western Armenian) is one of the most challenging questions for Armenian diachronic typology. As clefts have been documented since Classical Armenian, an evolution path can be drawn from phraseological clefts to grammaticalized auxiliary movement as a focus strategy in Eastern Armenian.

For characterizing clefts in Armenian, several parameters, such as agreement, prosody, type of connector, and the correlation between a cleft and its neutral counterpart, were outlined. Agreement in clefts involves two dimensions: intraclausal and interclausal agreement. As the empty pronoun is absent in Armenian, the impersonal feature is rendered by the lack of agreement between the subject and the copular verb in the matrix clause, allowing us to distinguish predicate clefts (impersonal with no agreement) from subject clefts (copular verb coindexed with personal pronouns as a subject). Our corpus-based data showed unequal representation of these two kinds of clefts (formally not distinguishable for 3SG subjects) depending on the person and the polarity, which is in line with common trends in cleft constructions cross-linguistically.

The neutral counterpart of a cleft in Armenian can be obtained by removing the copula and the complementizer, which is one of the main criteria for distinguishing clefts from other similar biclausal constructions.

In Armenian, constituents following the sentence stress (in our case, after the focused constituent) have some prosodic properties associated with backgrounding. Thus, if cleft constructions have the focused constituent at the beginning of a sentence, the rest of the sentence is in a backgrounded position. In the case of non-finite clefts, no such fixed order is required.

With the exception of the durative cleft construction in Eastern Armenian which can have inčh as a complementizer, the complementizer used in cleft constructions is exclusively the generic connector or “that”, which marks the beginning of the post-focus slot, characterized as backgrounded with a prosody of parenthesis.

Cleft constructions in Armenian mark a certain degree of focalization, parallel to other focus marking means. Focus marking in Armenian could be presented as a continuum rather than a sequence of clear-cut degrees. The scale of focalization depends on the kind of focus marking in Armenian, with simple prosody marking being the weakest, and preverbal position and clefts being the medium and the strongest ones respectively.

Unlike preverbal focus marking, which displays a number of important differences between Eastern Armenian and Western Armenian, the comparison of cleft constructions in the two standards has not shown real deviations. The existence of clefts with the same syntactic and pragmatic characteristics in Eastern Armenian and Western Armenian could constitute evidence that cleft constructions have diachronic continuity in Armenian, which makes the hypothesis of a grammaticalization path leading from cleft constructions to the auxiliary movement focus strategy more valid, though areal contact influence should not be neglected.

It was argued that cleft constructions mark a higher degree of emphasis, and that their distribution is discourse-dependent. As personal pronouns may be articulated either as clefted subjects or clefted predicates, we may discern a correlation between person and polarity and the preference for syntactic clefts. Since in Armenian we have on the one hand the syntactic biclausal cleft construction without an empty pronoun and on the other hand the auxiliary/ copula movement (monoclausal) strategy, the category of personal pronouns as clefted subjects occupies an intermediate position on this continuum.

As the coexistence of cleft constructions and other focus marking strategies (in-situ/preverbal, focus-sensitive particles) is typologically uncommon, it would be interesting to compare the data of Armenian with other languages with preverbal focus marking (especially those with the auxiliary movement strategy, e.g. with the Caucasian languages).

The present article being an attempt to draw a general picture of cleft constructions in Armenian, further research is necessary to deepen the analysis of such constructions from the semantic and pragmatic points of view, as well as to deepen the study of cleft-like constructions presented in this paper. Eastern and Western Armenian being the only two standard variants of Armenian, the integration of Armenian dialects into this research would be of great interest to complete the Modern Armenian continuum data, especially for checking the validity of the hypothesis of the grammaticalization path from cleft constructions to preverbal auxiliary/copula marking.


  • Abrahamyan S., 2004, Hajoch lezu. Šarahjusuthjun [The Armenian Language: Syntax], Yerevan.

  • Bybee J. et al. 1994, The evolution of grammar: Tense, aspect, and modality in the languages of the world, Chicago: The U. of Chicago Press.

    • Search Google Scholar
    • Export Citation
  • Bezrukov N. & Dolatian H., Affix order & information structure. Mobile affixes across Western Armenian. Talk presented at International Workshop of OV basic word order correlates and information structure, Paris, December 6–7, 2018.

    • Search Google Scholar
    • Export Citation
  • Comrie B., 1984, Some Formal Properties of Focus in Modern Eastern Armenian, Annual of Armenian Linguistics 5, p. 121.

  • Czypionka A., 2007, Word Order and Focus Position in the World’s Languages, Linguistische Berichte 212, p. 439454.

  • Donabedian A., 2010, Nom nu et tropisme typologique: le cas de l’arménien, in F. Floricic (ed.), Essais de typologie et de linguistique générale, mélanges Creissels, Lyon, ENS Editions, pp. 403416.

    • Search Google Scholar
    • Export Citation
  • Donabedian A., 2018a, Middle East and Beyond - Western Armenian at the crossroads: A sociolinguistic and typological sketch, in Christiane Bulut (ed.), Linguistic Minorities in Turkey and Turkic-speaking Minorities of the Peripheries: Turcologica 111, Harrassowitz Verlag, p. 89148.

    • Search Google Scholar
    • Export Citation
  • Donabedian A., 2018b, OV/VO variation in Western Armenian and IS correlates, presentation at International Workshop of OV basic word order correlates and information structure, Paris, 6–7 December, 2018.

    • Search Google Scholar
    • Export Citation
  • Donabedian A. & Sitaridou I., 2020, Anatolia, in A., Evangelia & Y. Matras, (eds.), The Routledge Handbook of Language Contact, Abingdon, Oxon: Routledge, p. 404433.

    • Search Google Scholar
    • Export Citation
  • Dum Tragut J., 2009, Armenian. Modern Eastern Armenian, Amsterdam and Philadelphia, John Benjamins.

  • Erschler D., 2012, From preverbal focus to preverbal “left periphery”: The Ossetic clause architecture in areal and diachronic perspective, Lingua 122, p. 673699.

    • Search Google Scholar
    • Export Citation
  • Forker D., 2021, Information structure in languages of the Caucasus, in M. Polinsky (ed.) The Oxford Handbook of Languages of the Caucasus, Oxford University Press, 9731003.

    • Search Google Scholar
    • Export Citation
  • Friedman V. A., 1996, Typological and Areal Features Linking and Separating the Balkans and the Caucasus, in J. Ian Press & F. E. Knowles (eds.), Papers in Slavonic Linguistics (Occasional Series) III, London, University of London, p. 99111.

    • Search Google Scholar
    • Export Citation
  • Gandon O., 2016, La relativisation dans une perspective aréale: l’aire Caucase – Anatolie de l’Est – Iran de l’Ouest. PhD thesis. Université Sorbonne Nouvelle, Paris.

    • Search Google Scholar
    • Export Citation
  • Göksel A. & Özsoy S., 2000, Is there a focus position in Turkish?, in A. Göksel & C. Kerslake (eds.), Studies on Turkish and Turkic Languages, Wiesbaden, Harrassowitz, p. 219228.

    • Search Google Scholar
    • Export Citation
  • Haig G., 2017, Western Asia: East Anatolia as a Transition Zone, in R. Hickey (ed.), The Cambridge Handbook of Areal Linguistics, Cambridge, Cambridge University Press, p. 396423.

    • Search Google Scholar
    • Export Citation
  • Harris A. C., 2001, Focus and universal principles governing simplification of cleft structures, in Jan T. Faarlund (ed.), Grammatical relations in change, Amsterdam and Philadelphia, John Benjamins, p. 159170.

    • Search Google Scholar
    • Export Citation
  • Harris A. C. & Campbell L., 1995, Historical Syntax in Cross-Linguistic Perspective, Cambridge, Cambridge University Press.

  • Hodgson K., 2019, Relative clauses in colloquial Armenian: Syntax and typology, PhD thesis, INALCO, Paris.

  • Jespersen O., 1927, A Modern English Grammar on Historical Principles, Part III, Syntax, vol.2, Heidelberg, C. Winter.

  • Kahnemuyipour A. & Megerdoomian K., 2017, On the Positional Distribution of an Armenian Auxiliary: Second-Position Clisis, Focus, and Phases, Syntax 20(1), p.7797.

    • Search Google Scholar
    • Export Citation
  • Kazenin K., 2002, Focus in Daghestanian and Word Order typology, Linguistic Typology 6, p. 289316.

  • Kiss K. É., 1987, Configurationality in Hungarian, Budapest, Akadémiai Kiado/ Dordrecht, D. Reidel.

  • Komen E. R., 2015, The Chechen it-cleft construction, Languages of the Caucasus 1(1), p. 76105.

  • Lambrecht K., K 1994, Information Structure and Sentence Form. Cambridge: Cambridge University Press

  • Lambrecht K., 2001, A framework for the analysis of cleft constructions. Linguistics 39, p. 463516.

  • Megerdoomian K., 2011, Focus and the Auxiliary in Eastern Armenian, paper presented at Berkeley Linguistic Society, February 2011, Languages of the Caucasus: Language, Gender, and Sexuality.

    • Search Google Scholar
    • Export Citation
  • Nicholas N., 1998, The story of pu: the grammaticalisation in space and time of a Modern Greek complementiser. PhD thesis, Department of Linguistics and Applied Linguistics, The University of Melbourne.

    • Search Google Scholar
    • Export Citation
  • Nichols J., 1994, Chechen, in R. Smeets (ed.), The indigenous languages of the Caucasus, vol. 4, The North East Caucasian languages II, Delmar, NY, Caravan Books, p. 177.

    • Search Google Scholar
    • Export Citation
  • Plungian V., 2018, Notes on Eastern Armenian verbal paradigms: “temporal mobility” and perfective stems, in D. Van Olmen, T. Mortelmans & F. Brisard (eds.), Aspects of linguistic variation: Studies in honor of Johan van der Auwera, Berlin, De Gruyter Mouton, p. 233245

    • Search Google Scholar
    • Export Citation
  • Samvelian P., Faghiri P. & Khurshudyan V., On the persistence of SVO: The case of Modern Eastern Armenian. Submitted.

  • Semenova X., 2014a, Механизмы фокусного маркирования в современном восточноармянском языке (The mechanisms of focus marking in Modern Eastern Armenian), in Труды Института лингвистических исследований РАН - Acta Linguistica Petropolitana, v. 10, # 1, Moscow, p. 838853.

    • Search Google Scholar
    • Export Citation
  • Semenova X., 2014b. Фокусное маркирование в современном восточноармянском языке «на дагестанском фоне» (Focus marking in Modern Eastern Armenian (on Dagestanian background)), in Язык. Константы. Переменные. Памяти Александра Евгеньевича Кибрика, Saint-Petersburg, p. 407429.

    • Search Google Scholar
    • Export Citation
  • Skopeteas S., Féry C. & Asatiani R., 2009, Word Order and Intonation in Georgian, Lingua 119(1), p. 102127.

  • Sornicola R., 2006, Interaction of syntactic and pragmatic factors on basic word order in the languages of Europe, in G. Bernini and M. Schwartz (eds), Pragmatic Organization of Discourse in the Languages of Europe, Berlin and New-York, Mouton de Gruyter, p. 357544.

    • Search Google Scholar
    • Export Citation
  • Stilo D. L., 2008, Two Sets of Mobile Verbal Person Agreement Markers in the Northern Talyshi Language, in S. Karimi et al. (eds), Aspects of Iranian Linguistics, Newcastle, Cambridge Scholars Publishing, p. 363390.

    • Search Google Scholar
    • Export Citation
  • Tamrazian A., 1991, Focus and Wh-Movement in Armenian, University College London Working Papers in Linguistics 3, p. 101121.


  •, accessed 5 December 2018: (Arak-29 Foundation).

  • Digital Library of Armenian Literature.

  • Eastern Armenian National Corpus.

  • Anaïd Donabedian’s personal corpus of Western Armenian processed by NooJ (press, literature, oral discourse)


We are grateful to Diana Forker, Katherine Hodgson, as well as to an anonymous reviewer for useful questions and comments.


The Armenian Kingdom of Cilicia (12th-14th centuries) officially favored abundant literature in vernacular Armenian, whereas no similar reality existed in Eastern Armenia.


The aorist, which remains a synthetic form in all variants of Armenian, is a notable exception, being the most stable verbal form in the whole Armenian diachronic and diatopic reality.


Abbreviations: gen genitive; pl plural; ipfv imperfective participle dat dative; do direct object; aux auxiliary verb; pfv perfective participle abl ablative; cop copula; dst1 destinative participle 1 ins instrumental; prs present; res resultative participle loc locative; pst past; sim simultaneous participle 1poss 1st person possessive; aor aorist; sbj subjective participle 2poss 2nd person possessive; cond conditional mood; conneg connegative participle 3poss 3rd person possessive; sbjv subjunctive mood; neg negation pl plural; imp imperative mood; proh prohibitive def definite article; med medio-passive; post postposition 1 1st person; caus causative; prox proximal 2 2nd person; inf infinitive; medl medial 3 3rd person; prog progressive; dist distal sg singular.


The transliteration system adopted in the paper is based on Hübschmann-Meillet system with certain modifications relevant in IPA system, and it does not reflect the phonetic differences between Western and Eastern Armenian standards.


According to EANC results, for the query with a bare object girkh kardal ‘to read a book’ with kardal ‘to read’ as a lexeme with all its paradigmatic forms and the direct object girkh ‘book’ in preverbal position, 615 matches have been found, whereas the query of the same phrase with the direct object in postverbal position had only 29 matches, thus we have 95% OV vs. 5% VO. The same verbal phrase with girkh ‘book’ having a definite/possessive article has 265 (≈ 70 %) matches for OV order versus 106 matches (≈ 30 %) of VO order, which shows a net distinction between the distribution of bare versus definite DOs (5% vs 30%).


In certain Armenian dialects (e.g. the Goris and Karabakh dialects) the canonical position of the auxiliary verb, be it affirmative or negative, is always after the converb, and the focus is marked by the inversion of the auxiliary to the post-focus position.


An additional piece of evidence for the enclitic status of the verb ē ‘to be’ is found in the distribution of the definite article. In Armenian, the suffixed definite article has two allomorphs: 1) –ǝ suffixed to a noun ending in a consonant; and 2) –n suffixed to a noun ending in a vowel. If a noun ending in a consonant is followed by a constituent with a vowel outset, both allomorphs (–ǝ and –n) are possible, with the suffixed –n being facultative for the cases where a separate intonational unit is possible. However, whenever the following constituent is a clitic (mainly the verb ē ‘to be’ and discursive markers), it is always pronounced as one prosodic unit, thus the optional –n becomes obligatory, as in the example below:



The movement of the verbal clitic k- with its allomorphs can be used in certain Western Armenian dialects to mark focus (Bezrukov & Dolatian 2018). This focus marking strategy bears a significant resemblance to the auxiliary/copula movement found in Eastern Armenian, and could provide evidence for some kind of grammaticalization continuity of language development in both Armenian branches.


Note that in Eastern Armenian, besides the movement of the auxiliary, the reverse order split between the auxiliary and the converb is possible as well (9d). The difference between (9c) and (9d) would be progressive focus ranging in the whole focus continuum of the statement.



The query ver+ kenal + ē in EANC results in only 3 matches (mainly questions with a verbal focus), whereas the same query with the verb in between the nominal element and the verb (i.e. ‘ver + ē + kenal’) gives 2215 matches. Similarly, the query girkh + kardal ‘to read’ + ē ‘to be’ has only 5 matches (again mainly the verb being questioned), whereas girkh book’ + ē ‘to be’ + kardal ‘to read’ has 294 matches.


Some general semantic constraints related to adverbial focus by auxiliary movement exist in MEA. For instance, tense modifiers (like arden ‘already’, deŕ ‘still’ etc.), which pertain to the assertion and not to the propositional content, cannot be focalized. Another interesting case that blocks auxiliary verb movement focus is the verbal phrase karoġ em ‘can’, which is originally composed of a non-finite form karoġ ‘capable’ and the copula ē ‘to be’.


The spectrum of uses of the complementizer or is particularly wide in Eastern Armenian oral discourse, where it can replace dedicated subordinate conjunctions of time (erb ‘when’), condition (ethe ‘if’), purpose (orpeszi ‘in order to’) (cf. Nicholas 1998) for similar development of the complementizer pu in Greek). This could account for or having an extremely high frequency of use and being the second most frequent conjunction in Armenian after ev ‘and’ (in EANC there are about 2 mln. tokens of or, i.e. over 1,8 % of the whole corpus, making it the third most frequent word after ē ‘is’ and ev ‘and’).


In the pair mišt ‘always’ / mišt ‘always’ + NEG, the negative form can be considered to be more salient than the affirmative one, which makes it more likely to appear in a CC than the affirmative one (cf. comparable results for the query “quantifier + angam ‘time’ + COP + or ‘that’ for which the matches with a negative copula are about four times more numerous than those with an affirmative one, which is accounted for by the expression It is not the first time that…”). The behavior of extreme time marking adverbs such as mišt ‘always’, erbekh ‘never’ (incompatible in a cleft construction) or hazvadep ‘rarely’ (only affirmative matches in cleft constructions) deserves further study.


The verb ē ‘to be’ is the main auxiliary verb in Armenian, whereas the verb linel ‘to be’ can function as an auxiliary verb in limited cases, e.g. in relative tense forms:



Thanks to Murad Suleymanov, Pollet Samvelian and Tabita Toparlak for Azeri, Persian and Turkish data respectively.


Cf. the statement in Harris (2001: 167) relevant also for Eastern Armenian: “The copula, which seems to be a reflex of the copula of the main clause of the cleft, seems at the same time to be involved as an auxiliary in the tense-aspect-mood system of some languages”.


This strategy is consistent with the preverbal focus strategy, since in (28a) replacing the periphrastic form by a synthetic one (e.g. the aorist: kančh-ech-in) would result in an OV order which would distinguish the neutral (nuclear stress) and contrastive foci not only formally but also by the intensity of prosodic prominence.

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 214 214 11
Full Text Views 5 5 0
PDF Views & Downloads 26 26 2