External Structure of Small Clauses in Modern Greek: The Case of Apodiknio and Anagnorizo
Author: Katerina Zoi, National and Kapodistrian University of Athens
Abstract: This paper explores the external structure of small clauses (SCs) in Modern Greek, focusing on the verbs apodiknio (prove) and anagnorizo (recognize). The first part sets the essential groundwork regarding the external structure of SCs, including Exceptional Case Marking, Control, and Raising. Apart from that, certain cases of verbs that appear to demonstrate similar structure are further explored. After establishing the necessary foundation, the second part of the paper turns to its main focus: the behavior of two Greek verbs, apodiknio and anagnorizo. Using a range of diagnostics, it is shown that both verbs exhibit structural and interpretive asymmetries. The findings contribute to our understanding of how SCs are licensed in Greek and how verb meaning interacts with syntactic environment.
Keywords: small clauses, Modern Greek, predication, apodiknio, anagnorizo, ECM, control, raising, HNP
1. Introduction
The term ‘small clauses’ (henceforth SCs) refers to structures that possess sentential characteristics, meaning that they contain a subject phrase and a predicative phrase ([DP XP], where XP can be an adjective, noun, or preposition) and typically serve as the complement of a verb. However, they differ from full sentences in that they are considered incomplete structures: They lack a verbal element, which is the prototypical predicate.
Two types of predicative configurations are commonly distinguished:
a) small clauses in which a linking verb intervenes [as in (1)], and
b) secondary predicative structures with no overt verbal element [as in (2)].
In (1), the small clause [DP XP] appears as the complement of a copula (ine)—leading to a Raising structure—whereas in (2), it is the complement of the verb theori (consider).
(1) I Antonia ine eksipni
the.NOM.SG.F Antonia be.3SG smart.NOM.SG.F
‘Antonia is smart.’
(2) I Katerina theori tin Antonia eksipni
the.NOM.SG.F Katerina consider.3SG the.ACC.SG.F Antonia smart.ACC.SG.F
‘Katerina considers Antonia smart.’
However, certain verbs exhibit unexpected restrictions when combining with small clause complements. For example, the verb apodiknio (prove) allows a small clause in the passive voice, but the active counterpart is not always acceptable for many speakers:
(3) To theorima apodichtike lathos
the.NOM.SG.N theorem prove.PST.PASS.3SG wrong.N.SG
‘The theorem turned out to be wrong.’
(4) ?[1]I Maria apedikse to theorima lathos
the.NOM.SG.F Maria prove.PST.ACT.3SG the.ACC.SG.N theorem wrong.N.SG
‘Maria proved the theorem wrong.’
This asymmetry raises the following research question: What structural and interpretive factors constrain the licensing of small clauses in Greek, and why are these constraints sensitive to the choice of verb and voice?
This paper addresses this question by examining the external structure of small clauses and the behavior of two Greek verbs—apodiknio (prove) and anagnorizo (recognize)—which exhibit similar asymmetries. The aim is to determine the reason: Do they arise from the voice or the Aktionsart of the verbs involved?
Section 2 provides the necessary background by surveying small clause configurations in Greek, including Exceptional Case Marking (ECM), Control, and Raising. This structural overview forms the foundation for the second part of the paper, which focuses on the syntactic and interpretive properties of apodiknio and anagnorizo. Using a range of diagnostics, I argue that their acceptability depends not on voice alternation per se, but rather on the Aktionsart of the verb.
2. The External Structure of Small Clauses
To determine the relationship between small clauses (SCs) and the matrix verb, the first step is to investigate how the subject of the SC connects with it. If there is Exceptional Case Marking (ECM), the DP of the SC functions as its subject, receiving a thematic role from the predicate of the SC and accusative case from the sentence verb. In contrast, if the construction involves Control, the DP becomes a complement of the verb and receives both a thematic role and case from it. In that case, the SC’s subject is a null pronominal (pro), co-indexed with the matrix DP. As this paper focuses on Modern Greek, a pro-drop language, pro is used to capture null predicative subjects that are recoverable via agreement or co-reference. One more structure that must be considered involves raising: when the subject of the SC raises to [Spec, IP]. Finally, entire SCs may appear in subject position of the matrix clause (Honorary Noun Phrases).
2.1 ECM Structures
In Exceptional Case Marking (ECM) structures, the entire SC functions as the complement of the verb. Consider the following example:
(5) Theoro [SC ton Robbie Williams omorfo]
consider.1SG the.ACC.SG.M Robbie Williams handsome.ACC.SG.M
‘I consider Robbie Williams handsome.’
Here, the SC ton Robbie Williams omorfo is the complement of theoro (consider). The DP ton Robbie Williams receives a thematic role from the AP omorfo (handsome) and accusative case from the matrix verb. Since adjectives cannot assign case, the case is assigned by the verb. The adjective receives the same case, as in Greek the noun and the adjective must carry the same case, number, and gender (when applicable).
2.2 Control Structures
When the DP is an argument of the matrix verb and receives a thematic role from it, the small clause occupies an adjunct position. The subject of the SC is a null pronominal (pro), co-indexed with the matrix DP. In these cases, the SC expresses a secondary predication about that DP.
(6) I Dimitra cheretise ton Kostaᵢ [SC proᵢ proto]
the.NOM.SG.F Dimitra.NOM greet.PST.3SG the.ACC.SG.M Kostas.ACC first.ACC.SG.M
‘Dimitra greeted Kostas first.’
(7) I Dimitraᵢ cheretise ton Kosta [SC proᵢ proti]
the.NOM.SG.F Dimitra.NOM greet.PST.3SG the.ACC.SG.M Kostas first.NOM.SG.F
‘Dimitra was the first to greet Kostas.’
(8) I Dimitraᵢ perpatuse [SC proᵢ skythropi]
the.NOM.SG.F Dimitra.NOM walk.PST.3SG gloomy.NOM.SG.F
‘Dimitra was walking gloomily.’
(9) I Dimitraᵢ irthe [SC proᵢ proti]
the.NOM.SG.F Dimitra.NOM come.PST.3SG first.NOM.SG.F
‘Dimitra came first.’
In these cases, the SC serves as an adjunct expressing a secondary property or manner. The null subject pro is licensed and interpreted via co-indexation with one of the verb’s arguments [subject in examples (6), (8), (9) or object in example (7)].
These predicative adjuncts occur across different verb classes:
Transitive verbs: (6), (7)
Unergative verbs: (8)
Unaccusative verbs: (9)
Even when transitive verbs form passive constructions, the subject of the SC is still controlled by an internal argument. The predicate is in nominative case:
(10) I Nektariaᵢ eklechthike [SC proᵢ proedros]
the.NOM.SG.F Nektaria elect.PST.PASS.3SG president.NOM.SG.M
‘Nektaria was elected president.’
2.3 Control or ECM?
2.3.1 Quasi-ECM Cases
In certain cases, it is unclear whether the DP belongs to the SC or is a direct complement of the matrix verb. This ambiguity appears in quasi-ECM structures with verbs like thelo (want) and perimeno (expect), which exhibit properties of both ECM and control.
The following examples illustrate the syntactic ambiguity in question:
(11) Thelo ton Giorgo na pari luludia sti gineka tu
want.1SG the.ACC.SG.M Giorgos.ACC.SG SBJV take.3SG flowers to.the.SG.F wife his
‘I want Giorgos to buy flowers for his wife.’
(12) Thelo o Giorgos na pari luludia sti gineka tu
want.1SG the.NOM.SG.M Giorgos.NOM SBJV take.3SG flowers to.the.SG.F wife his
‘I want Giorgos to buy flowers for his wife.’
(13) Thelo ton Giorgo etimo gia avrio
want.1SG the.ACC.SG.M Giorgos.ACC ready.ACC.SG.M for tomorrow
‘I want Giorgos ready for tomorrow.’
According to Kotzoglou (2002), verbs like thelo and perimeno do not form ECM structures. Both the DP, when placed in the accusative case, and the CP function as complements of the verb in the main clause, and therefore the DP is part of the main clause, not the subordinate one. Assuming this, the DP is a complement to the verb, and the SC has a pro as its subject, while functioning as the second complement of the verb.
Having the above in mind, it should be further investigated what the correct structure is and whether this represents a Control structure [examples (14) and (15)] or ECM [examples (16) and (17)]:
(14) Thelo [DP ton Giannii] [SC proi nekro]
want.1SG the.ACC.SG.M Giannis.ACC pro dead.ACC.SG.M
‘I want Giannis dead.’
(15) Perimeno [DP ton Giannii] [SC proi etimo]
expect.1SG the.ACC.SG.M Giannis.ACC pro ready.ACC.SG.M
‘I expect Giannis to be ready.’
(16) Thelo [SC ton Gianni nekro]
want.1SG the.ACC.SG.M Giannis.ACC dead.ACC.SG.M
‘I want Giannis dead.’
(17) Perimeno [SC ton Gianni etimo]
expect.1SG the.ACC.SG.M Giannis.ACC ready.ACC.SG.M
‘I expect Giannis to be ready.’
To determine this, certain diagnostics, such as those mentioned in Kotzoglou (2002: 45-49), must be applied. First, we add the light verb ‘be,’ which does not carry any specific semantic weight. Secondly, by applying negative polarity items, it is observed that they can only appear in nominative case when the negation concerns the subordinate clause. If they appear in accusative, the structure is ungrammatical:
(18a) I Maria ithele kanenas na min ine skithropos
the.NOM.SG.F Maria want.PST.3SG no.one.NOM.SG.M SBJV not be.3SG gloomy.NOM.SG.M
‘Maria wanted no one to be gloomy.’
(18b) *I Maria ithele kanenan na min ine skithropos
the.NOM.SG.F Maria want.PST.3SG no.one.ACC.SG.M SBJV not be.3SG gloomy.ACC.SG.M
Intended: ‘Maria wanted no one to be gloomy.’
The same happens with the pronoun i idia (herself):
(19a) I Sevii ithele i idiai na ine i daskala tus
the.NOM.SG.F Sevi.NOM want.PST.3SG the.NOM.SG.F self.NOM.SG.F SBJV be.3SG the teacher their
‘Sevi wanted herself to be their teacher.’
(19b) *I Sevii ithele tin idiai na ine i daskala tus
the.NOM.SG.F Sevi.NOM want.PST.3SG the.ACC.SG.F self.ACC.SG.F SBJV be.3SG the teacher their
Intended: ‘Sevi wanted herself to be their teacher.’
Even in the case of idiomatic expressions, the DP can only be in the nominative case. The subjects of idiomatic expressions cannot receive any thematic role other than the one they take within the expression itself. Therefore, if the examined verbs take only the DP as a complement, and not the entire SC, the structures that are formed exclude the semantic interpretation of the SC as an idiomatic expression. The idiom to be used is the Greek expression i mana mu mes sta podia mu (i.e., my mother is interfering with my personal matters, literally meaning ‘my mother is between my legs’):
(20a) O Dimos den ithele i mana tu na ine mes sta podia tu
the.NOM.SG.M Dimos not want.PST.3SG the.NOM.SG.F mother.NOM his SBJV be.3SG in.the legs his
(20b) *O Dímos den ithele ti mana tu na ine mesa sta podia tu
the.NOM.SG.M Dimos not want.PST.3SG the.ACC.SG.F mother.ACC his SBJV be.3SG in.the legs his
According to the above criteria, what holds for every subordinate clause also holds for SCs with a light verb. This is, of course, expected due to the presence of the complementizer na. The cases examined here pertain to subordinate clauses and therefore cannot serve as evidence for what really happens with SCs. When the light verb and complementizer are absent, while there are no verbal characteristics inside the SC, and the idiomatic criterion is applied, the following is observed:
(21) Den thelo ti mana mu mes sta podia mu
not want.1SG the.ACC.SG.F mother.ACC my in.the legs my
(22) Den perimena ti mana mu mes sta podia mu
not expect.PST.1SG the.ACC.SG.F mother my in.the legs my
From the above, it is evident that the idiomatic expression holds its interpretation and its meaning is not altered. The DP receives a thematic role from the SC and not from the matrix verb. The entire SC functions as a complement to the verb.
It is also worth noting that the same ECM structure is observed with the verb echo (have)
(23) Eho [SC ton Gianni arrosto]
have.1SG the.ACC.SG.M Giannis.ACC sick.ACC.SG.M
‘I have Giannis sick.’
If the SC ton Gianni arrosto is replaced by an idiomatic expression like to kefali mu isicho (meaning, I am not worried, literally ‘my head calm’), a grammatical result is produced where the idiomatic expression retains its meaning [example (24)]:
(24) Echo to kefali mu isicho
have.1SG the.ACC.SG.N head.ACC my calm.ACC.SG.N
It is also worth mentioning that SCs can function as complements in prepositional phrases as well:
(25) Me [SC to pedi arrosto], den boro na figo apo to spiti.
with the.ACC.SG.N child.ACC sick.ACC.SG.N not can.1SG SBJV leave.1SG from the.ACC.SG.N house
‘With the child sick, I can’t leave the house.’
In example (25), the DP to pedi (child) receives a thematic role from the predicate arrosto (sick) and case from the preposition me (with). This is also an ECM case. The preposition, governing the DP, assigns case, and the predicate then receives the same case.
2.3.2 Control Cases
Predicative small clauses also appear with perception verbs, such as vlepo (see) and akouo (hear):
(26) Ida [DP tin Ntinaᵢ] [SC proᵢ chamogelasti]
see.PST.1SG the.ACC.SG.F Ntina.ACC smiling.ACC.SG.F
‘I saw Ntina smiling.’
Example (26) differs from other Control Structures, because the SC here does not function as an adjunct but as a complement of the verb. In other words, the verb in the sentence has two objects, the DP and the SC. It also differs from verbs with quasi-ECM structures, because here the DP receives a thematic role from the verb of the sentence, not from the predicate of the SC. This can be easily verified by applying the same idiom test used earlier:
(27) Den ida ti mana mu mes sta podia mu
not see.PST.1SG the.ACC.SG.F mother.ACC my in.the legs my
Indeed, the idiom loses its meaning, leading to the conclusion that this is a Control Structure.
2.4 Raising Structures
SCs can be the only argument of some unaccusative verbs, primarily ime (be) and ginome (become):
(28) Ine [SC i Nansi eksipni]
be.3SG the.NOM.SG.F Nansi smart
‘Nancy is smart.’
In example (28), the DP subject of the SC has not received a case because, as mentioned earlier, the predicate eksipni (smart) cannot assign one. Thus, the DP Nansi raises to the [Spec, IP] position of the verb in the sentence, receiving a nominative case. Then, the predicate appears also in nominative:
(28’) I Nansi ine eksipni
the.NOM.SG.F Nansi.NOM be.3SG smart.NOM.SG.F
‘Nancy is smart.’
The difference between the aforementioned unaccusative verbs, and others like in example (9) lies in the fact that in the first case, the argument of the sentence verb is the entire SC, whereas in the second case, it is the DP, while the SC functions as an adjunct, with its subject being controlled by the verb’s complement.
A special case involves certain verbs, which, although they seemingly do not accept an SC in the active voice, do accept one in the passive:
(29) O Iakovos apedikse to theorima
the.NOM.SG.M Iakovos.NOM prove.PST.3SG the.ACC.SG.N theorem
‘Iakovos proved the theorem.’
(30) To theorima apodichtike lathos
the.NOM.SG.N theorem.NOM.SG.N prove.PST.PASS.3SG wrong.SG.N
The theorem was proven wrong.’
(31) */?O Iakovos apedikse to theorima lathos
the.NOM.SG.M Iakovos prove.PST.3SG the.ACC.SG.N theorem.ACC.SS.N wrong.SG.N
Intended: ‘Iakovos proved the theorem wrong.’
Another structure that is worth mentioning is Honorary Noun Phrases. Honorary Noun Phrases (HNPs) refer to small clauses that appear in subject position. In Greek, they are relatively rare and often marked in terms of information structure or focus. Their syntactic behavior diverges from canonical subject DPs, and their internal predicative relationship resembles that of typical small clauses. Although not central to the present analysis, HNPs are mentioned here for completeness.
Now that we have surveyed the external structure of small clauses in Modern Greek—covering ECM, Control, and Raising—we are in a position to examine the verbs that exhibit non-standard behavior and are mentioned before. The next section focuses on apodiknio (prove) and anagnorizo (recognize), whose interaction with small clause complements raises intriguing questions about argument structure and Aktionsart.
3. Syntax of Small Clauses with the Verbs apodiknio (prove) and anagnorizo (recognize)
Some verbs in Greek exhibit the odd phenomenon that, while they generate grammatical sentences in passive voice, when combined with a small clause (SC) in active voice, they may produce ungrammatical sentences. These verbs are mainly apodiknio (prove) and anagnorizo (recognize). The following examples illustrate the contrast between passive and active uses of apodiknio in SC environments. It is worth mentioning that there is an inconsistency regarding whether certain sentences are acceptable or not to native Greek speakers. Some consider example (33) ungrammatical, while others do not find anything wrong.
(32) To theorima apodichthike lathos
the.NOM.SG.N theorem prove.PST.PASS.3SG wrong.SG.N
‘The theorem was proved to be wrong.’
(33) ?O Giannis apedikse to theorima lathos
the.NOM.SG.M Giannis prove.PST.3SG the.ACC.SG.N theorem.ACC wrong.SG.N
‘Giannis proved the theorem wrong.’
(34) O Giannis apodichthike megalos apateonas
the.NOM.SG.M Giannis prove.PST.PASS.3SG big.NOM.SG.M fraudster.NOM.SG.M
‘Giannis turned out to be a big fraud.’
(35) *I María apedikse ton Gianni megalo apateona
the.NOM.SG.F Maria prove.PST.3SG the.ACC.SG.M Giannis.ACC big.ACC.SG.M fraudster.ACC.SG.M
Intended: ‘Maria proved Giannis a big fraud.’
(36) O Giannis apedikse to theorima
the.NOM.SG.M Giannis prove.PST.3SG the.ACC.SG.N theorem
‘Giannis proved the theorem.’
(37) To theorima apodichthike
the.NOM.SG.N theorem prove.PST.PASS.3SG
‘The theorem was proven.’
As mentioned before, some native speakers find (33) acceptable. However, only the passive forms [examples (32), (34), (37)] are consistently accepted by all native speakers. In contrast, (35) is not acceptable by any native Greek speaker. Structures might be acceptable in cases where the subject of the SC involves the notion of “provability,” but when the subject does not carry this meaning, no structure is acceptable. A theorem can be proved, but a person cannot.
The aforementioned examples demonstrate that apodiknio may select a small clause complement in the passive, but in the active voice, the element that appears to be the subject of the SC is actually the complement of the verb, and the SC itself functions as an adjunct:
(33’) O Giannis apedikse to theorimaᵢ [SC proᵢ lathos]
the.NOM.SG.M Giannis prove.PST.3SG the.ACC.SG.N theorem.ACC wrong.SG.N
‘Giannis proved the theorem wrong.’
Structurally, apodiknio appears to allow two structures: A) a transitive one [example (36)], where the verb has both an external and an internal argument, and B) an unaccusative [examples (32), (33)], where a SC may function as a single internal argument. Although example (32) is a passive and not a true unaccusative, the argument structure aligns with an unaccusative-like.
Why does this happen? Is apodiknio an ergative verb with two structures? Or does the meaning of the verb change, altering its Aktionsart? If we consider examples (32)-(33) and (36)-(37), it seems there is no change in meaning: Indeed, apodiknio is a verb that belongs to the category of achievements [according to Vendler’s (1957) classification]. When someone proves something, like a theorem, or when a theorem is proved, or even when a theorem is proven false, or when someone proves a theorem false, it is an instantaneous event that occurs at a specific point in time. Let’s observe the following examples, exploring a certain meaning of the verb:
(38) *O Giannis apodikniei / apodiknieye / apedikse to theorima mia ora
the.NOM.SG.M Giannis prove.IMPFV.3SG / prove.PST.IMPFV.3SG / prove.PST.PFV.3SG the.ACC.SG.N theorem one hour
Intended: ‘Giannis proves / was proving / proved the theorem for an hour.’
(39) *To theorima apodikniete / apodikniotan / apodichthike mia ora
the.NOM.SG.N theorem prove.IMPFV.PRS.PASS.3SG / prove.IMPFV.PST.PASS.3SG / prove.PST.PFV.PASS.3SG one hour
Intended: ‘The theorem is being proved / was being proved / was proven for an hour.’
(40) *To theorima apodikniete / apodikniotan / apodichthike lathos mia ora
the.NOM.SG.N theorem prove.IMPFV.PRS.PASS.3SG / prove.IMPFV.PST.PASS.3SG / prove.PST.PFV.PASS.3SG wrong.SG.N one hour
Intended: ‘The theorem is being proved / was being proved / was proven wrong for an hour.’
(41) *O Giannis apodikniei / apodiknieye / apedikse to theorima lathos mia ora
the.NOM.SG.M Giannis prove.PRS.3SG / prove.IMPFV.PST.3SG / prove.PFV.PST.3SG the.ACC.SG.N theorem wrong.SG.N one hour
Intended: ‘Giannis proves / was proving / proved the theorem wrong for an hour.’
(42) O Giannis apedikse / *apodiknieye to theorima (lathos) mesa se mia ora
the.NOM.SG.M Giannis prove.PFV.PST.3SG / *prove.IMPFV.PST.3SG the.ACC.SG.N theorem wrong.SG.N in one hour
‘Giannis proved/ *was proving the theorem (wrong) in one hour.’
(43) To theorima apodichthike / *apodikniotan (?lathos) mesa se mia ora
the.NOM.SG.N theorem prove.PFV.PST.PASS.3SG / prove.IMPFV.PST.PASS.3SG wrong.SG.N in one hour
‘The theorem was proven / *was being proven (wrong) in one hour.’
From the above, one can conclude the following: A theorem is proved or proved false in a moment, and from that point, this does not change. These data indicate that the verb behaves like an achievement verb (Vendler, 1957) when both external and internal arguments are present. However, there are examples like:
(44) I Maria apodichthike poli kali fili ola afta ta chronia
the.NOM.SG.F Maria prove.PFV.PST.PASS.3SG very good.NOM.SG.F friend.NOM.SG.F all these the years
‘Maria turned out to be a very good friend all these years.’
(45) *I Maria apodichthike poli kali fili mesa se enan chrono
the.NOM.SG.F Maria prove.PFV.PST.PASS.3SG very good.NOM.SG.F friend.NOM.SG.F in one.ACC.SG.M year
Intended: ‘Maria turned out to be a very good friend in one year.’
The verb no longer denotes an achievement, but rather an activity. Maria becomes a good friend over time and may continue to be a good friend for some time too. The fact that Maria is proved to be a good friend does not imply that she is permanently “proven” as such; she may also stop being. In these cases, apodiknio aligns more with an activity.
Thus, the contrast is not due to voice per se, but to the argument structure: when the verb selects both external and internal arguments, it is an achievement. When only the internal argument is present (as in passive or unaccussative structures), it aligns with activities.
The objection someone might raise concerns whether apodiknio in the passive voice is truly an achievement when it has an SC as a complement, and whether this is true in the case where the subject of the SC carries the meaning of provability, as in example (33). It is not a coincidence, though, that such sentences may be accepted in active voice, unlike example (35): to prove a theorem false, one must perform the proof, whereas when someone is proved to be a good friend, this is not done by anyone else. Therefore, in the first case, there is an AGENT, whereas in the second, there is not. This fact is indeed a very strong argument regarding the above.
In other words, the difference in voice is the result and not the cause. It is not the voice that determines the meaning of the verb, but the difference in meaning is reflected in the voice. When apodiknio is an activity and not an achievement—and is unaccussative—it must be in passive voice, where the internal argument is located in the subject position.
The same Aktionsart alternation appears with anagnorizo (recognize):
(46) I Maria anagnorise ton Gianni
the.NOM.SG.F Maria recognize.PST.3SG the.ACC.SG.M Giannis.ACC
‘Maria recognized Giannis.’
(47) O Giannis anagnoristike (apo ti Maria)
the.NOM.SG.M Giannis.NOM recognize.PST.PASS.3SG (by the.ACC.SG.F Maria)
‘Giannis was recognized (by Maria).’
(48) O Giannis anagnoristike os to simantikotero prosopo tis telefteas dekaetias
the.NOM.SG.M Giannis recognize.PST.PASS.3SG as the.NOM.SG.N most.important.NOM.SG.N person.NOM.SG.N of.the last decade
‘Giannis was recognized as the most important person of the last decade.’
(49) ?I Maria anagnorise ton Gianni os to simantikotero prosopo tis telefteas dekaetias
the.NOM.SG.F Maria.NOM recognize.PST.3SG the.ACC.SG.M Giannis.ACC as the.ACC.SG.N most.important.ACC.SG.N person.ACC.SG.N of.the last decade
‘Maria recognized Giannis as the most important person of the last decade.’
Here, the SC may not be accepted as a complement in the active voice. The word “may” again suggests that this observation is not totally ungrammatical.
Same with apodiknio, it is the Aktionsart of the verb that changes in anagnorizo: In examples (46)-(47), the transitive verb represents an achievement, as it is instantaneous, without duration. On the contrary, in (48), the unaccussative verb is actually an activity, as it extends over time; however, it does not extend between specific temporal limits, nor does it necessarily result in a definite outcome. Thus, while Giannis is recognized by Maria and from then on is considered recognized by her, when he is recognized as the most important person of the decade, it is not the case that he is permanently “recognized,” as this process extends over time and may stop at any moment.
4. Conclusion
Small clauses in Modern Greek can be combined with all types of verbs, depending on their position in the sentence and how their subject is assigned case. When the SC is a complement of the verb, its subject receives a thematic role from the predicate of the SC and an accusative case from the verb of the sentence (ECM structures). Such structures are formed by transitive verbs. When the SC is an adjunct to the verb, its subject is a pro and not the direct object (Control structures). Such structures can be formed by all types of verbs. In Raising structures, the subject of the SC receives a thematic role from the predicate of the SC and is raised to the subject position of the verb, where it receives the nominative case. Such structures are formed by intransitive verbs and verbs in the passive voice. Additionally, SCs can also sometimes be placed in the subject position (HNPs).
There are some verbs in Greek that exhibit peculiarities in accepting an SC as a complement. Although voice or syntax may seem to play a role, it is actually the argument structure of the verb that influences the distribution. Different argument structures imply different Aktionsart, thus sometimes leading to different types of complements.
Bibliography:
Aarts, B. (1992). Small clauses in English (Topics in English Linguistics, Vol. 8). Berlin: Mouton de Gruyter.
Chomsky, N. (1986b). Barriers. Cambridge, MA: MIT Press.
Kotzoglou, G. (2002). Greek ‘ECM’ and how to control it. Reading Working Papers in Linguistics, 6, 39–56.
Nishihara, T. (2003). Equative construction with honorary NPs in English. Nagasaki University English Linguistics, 20(2), 395–418.
Safir, K. (1983). On small clauses as constituents. Linguistic Inquiry, 14(4), 730–735.
Spyropoulos, V. (1999). Agreement relations in Greek. Reading: University of Reading.
Spyropoulos, V., & Stamatogiannis, N. (2011). Subextraction from subjects in Greek. Paper presented at the conference Islands in Contemporary Linguistic Theory, University of the Basque Country, Vitoria-Gasteiz, Spain.
Stowell, T. (1991). Small clause restructuring. In R. Freidin (Ed.), Principles and parameters in comparative grammar (pp. 182–219). Cambridge, MA/London, England: MIT Press.
Tsokoglou, A. (1995). Small clauses in Modern Greek. Paper presented at the 2nd International Conference on Greek Linguistics, Salzburg, Austria.
Vendler, Z. (1957). Verbs and times. The Philosophical Review 66 (2), 143–160.
Williams, E. (1980). Predication. Linguistic Inquiry 11 (1), 203–238.
Williams, E. (1983). Against small clauses. Linguistic Inquiry 14 (2), 287–308.
[1] Judgment notation: * indicates ungrammaticality, ? indicates marginal acceptability, and ?? indicates strong degradation. Grammatical examples are unmarked. Judgments are based on the author’s native speaker intuitions and have been cross-checked where relevant.