Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS (linguistics.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 23 June 2017

Passives and Syntax

Summary and Keywords

Cross-linguistic differences in passive formation and the differences between verbal and adjectival passives reveal some of the core properties of the passive. In earlier stages of the Principles and Parameters framework, differences in both these domains were taken as evidence that the grammar has two distinct components to build passives, namely the lexicon and the syntax. This intuition can be restated by adopting the view that all passive formation is syntactic. Indeed, it has been posited that there are two syntactic domains to build passives, and these two domains correlate with distinct properties of passive formations within a language and across languages.

Keywords: passive, case absorption, adjectival passives, middle

1 Introduction

The passive is one type of a transitivity alternation that a particular verb may enter in a language. Consider for example the data in (1):

(1)

Passives and Syntax

As we see in (1), the active differs from the passive variant in that in (1a) the external argument of the active predicate, i.e., John, also bears the grammatical function of subject; in (1b), however, it is the verb’s internal argument, i.e., the window that resumes this grammatical function. The external argument of the transitive predicate may be expressed via a prepositional phrase (PP), e.g., the English by. In the absence of a by-phrase, the passive is referred to as short or basic. However, both examples have a very similar interpretation: they both entail that someone broke the window, i.e., the external argument of the active predicate is present implicitly, even when it is not expressed via a PP. The difference in grammatical functions observed in (1) is found in other transitivity alternations such as the causative alternation; however, the presence of an implicit external argument is not shared by causative alternation, as we will see below. There has been a lot of discussion in the literature as to whether or not the external argument in passive sentences is syntactically accessible.

A second difference between the passive and the active is formal: as we see in English, passive formation involves the presence of the auxiliary be in combination with the participle. This type of passive is called periphrastic or analytic. Other languages such as Hebrew and Greek form synthetic passives, i.e., the form of the active verb changes. In Greek, this includes the presence of the so-called nonactive affix (NAct):

(2)

Passives and Syntax

Haspelmath (1990) actually claimed that there is no language that marks passive and active alike, but see Pitteroff (2015) for some arguments against this view.

According to Keenan and Dryer (2006), not all of the world’s languages have passives, but if a language has a passive with an agent/by-phrase, then it will also have a basic/short passive. In this article, I will not be concerned with the cross-linguistic availability of the passive; rather I will concentrate on languages that have passives and I will touch upon the following two issues. First of all, it has been pointed out in the literature that passives are not identical across languages (see also Huang, 1999). While in earlier and some later approaches this was captured in terms of lexical vs. syntactic passive formation (see, e.g., Reinhart & Siloni, 2005), some of the recent literature on this topic argues that it is the syntactic construction of passives that differs across languages (see e.g., Alexiadou & Doron, 2012; Doron, 2015; Alexiadou, Anagnostopoulou, & Schäfer, 2015). A second, related issue concerns the differences between verbal and adjectival passives. Again, while earlier literature, most prominently Wasow (1977), treated these differences as the result of a syntactic (for verbal passives) vs. lexical (for adjectival passives) derivation, more recent literature (e.g., Embick, 2004; Anagnostopoulou, 2003; Kratzer, 2000 and others building on their results) argues that both are syntactic and they only differ in terms of their syntactic make-up.

A third and again more recent controversy in the discussion of adjectival passives concerns the presence of an implicit external argument in them. Several authors have recently argued that even in languages such as English and German, adjectival passives may contain implicit external arguments. Thus the presence of the structural layer that introduces external arguments seems not to be parametrized, as argued for in Anagnostopoulou (2003). This, however, raises the question what the real differences are between verbal and adjectival passives, if any. In view of the fact that we by now have very good descriptions of the empirical domain, a detailed overview of these issues can be offered, which hopefully will prompt further discussion.

The article is structured as follows. I will first briefly summarize the properties we typically associate with passives and the diagnostics used in the literature to provide evidence for the presence of an implicit external argument in them. I will then turn to issues of cross-linguistic differences with respect to verbal passive formation and then address the issue of verbal vs. adjectival passives.

2 How Are Passives Derived?

The relationship between active and passive sentences has been a constant preoccupation for linguists working in various theoretical models. Within the generative grammar framework, the first proposal concerning the structure of the passive is found in Chomsky (1957). Chomsky proposes an optional transformational rule that inserts the auxiliary be as well as the affix -en to the main verb and leads to a reordering of the noun phrases involved as in (3). Chomsky notes that this rule needs to take a transitive input. In fact, for every transitive sentence in English there is the corresponding passive one. This in turn means that the passive structure is derived from the active one.

(3)

Passives and Syntax

In Chomsky (1965), a slightly more general rule is employed. Nevertheless, the surface subject is analyzed as the underlying object of the verb that then raises in the passive structure.

In Chomsky (1981), the perspective changes in that one no longer makes reference to a specific passive transformation. Rather passive is one of the structures best analyzed under a more general rule, namely Move α‎. Specifically, the passive involves NP movement of the internal argument of the predicate to subject position, as illustrated in (4), an analysis that also applies to, e.g., unaccusatives:

(4)

Passives and Syntax

This movement is motivated as follows: the passive verb fails to assign accusative Case to its internal argument; as a result, this noun phrase has to move to a position where it can be assigned Case. The external argument of the verb is absorbed. Thus passive verbs fail both to assign accusative Case and they also lack an external argument, an analysis that also applies to unaccusative predicates. The correlation of these two properties is referred to in the literature as Burzio’s generalization (Burzio, 1986). Baker, Johnson, and Roberts (1989) propose that it is actually the passive suffix that ‘absorbs’ the accusative Case and external theta-role of the verb. The direct object (lacking Case in its VP internal position) then raises to Spec,T/IP, where it can get nominative Case satisfying the Case filter.

However, it is controversial whether the case absorption feature is a universal characteristic of the passive, as there are several languages where accusative Case seems to be retained in the passive variant. For instance, Åfarli (1992) suggests that this is possible in Norwegian, Sobin (1985) claims the same for Ukranian; see also Lavine (2013) and Eythórsson (2008) for Icelandic, but see Maling and Sigurjónsdóttir (2002) for an alternative analysis. See Huang (1999) for a general discussion on this issue.

In Chomsky (1981), the assumption is that the single NP argument in the passive is generated in its thematic position at the level of D(eep)-Structure and moves to its surface position at the level of Surface-Structure. However, unlike in earlier work, the passive is not derived from the active, but both active and passive versions of a sentence have distinct D-structures.

In frameworks such as relational grammar use is made of different strata where properties related to subject and object are appropriated. From this perspective, the differences in word order and agrement between active and passive sentences are attributed to the fact that the object of the active sentence role has been advanced to the subject position and the subject of the active sentence is demoted to the framework specific role of the chommeur. This takes place in a second stratum, which is in addition to the first one where the active sentence is formed (see Perlmutter & Postal, 1983).

In lexical frameworks, such as lexical functional grammar (Bresnan, 1982), passivization involves both a functional change and a morphological change, but crucially these are properties that are specified in the lexicon by means of lexical rules.

Kiparsky (2013) provides an analysis of the passive within the lexical decomposition model and claims that the syntax of the passive ‘is predictable from the language’s active sentences and the argument structure of passive predicates, which is derived from the argument structure of the basic predicate by an invariant operation triggered by the passive morpheme.’ See also Wunderlich (1993) and Müller (2007) for alternative lexical approaches to the passive. See Reinhart and Siloni (2005), who argue that operations on argument structure can take place either in the syntax or in the lexicon, and this is subject to cross-linguistic variation.

3 Diagnosing Implicit External Arguments

A number of tests have been proposed in the literature in order to diagnose the presence of an implicit external argument in the passive (see, e.g., Baker, Johnson, & Roberts, 1989; Bhatt & Pancheva, 2006). These diagnostics show that, unlike in other transitivity alternations such as the causative alternation, an implicit external argument is present in the passive. By contrast, anticausatives, i.e., intransitive variants of predicates entering the causative alternation, lack such an argument. The first diagnostic is the licensing of by-phrases in languages that have such phrases. As we saw above, and is also illustrated in (5c), the passive allows modification by a by-phrase, while this is not the case for anticausative verbs (5b).

(5)

Passives and Syntax

The explanation for this is that only the passive contains an implicit external argument, which in turn can license the by-phrase.

The second observation is that the implicit external argument of a passive is able to control the PRO subject of a rationale clause. Since anticausatives lack such an argument, control is impossible. This is shown by the well-known example in (6):

(6)

Passives and Syntax

Another test involves the availability of agentive adverbs such as deliberately. These can appear in passives and relate to the implicit agent, which is not the grammatical subject of the passive. In the anticausative case, however, there is no implicit agent, hence the adverb necessarily associates with the grammatical subject, and as a result the sentence is deviant.

(7)

Passives and Syntax

A fourth environment is the case of instrumental PPs. The consensus is that such PPs must be in the control of an external argument, and thus can only appear in passives but not in anticausatives.

(8)

Passives and Syntax

Baker, Johnson, and Roberts (1989) discuss a further test diagnosing the presence of a verb’s external argument in verbal passives, namely the fact that verbal passives do not allow a self-action interpretation of the verb. In this, verbal passives differ from adjectival passives, which allow a self-action interpretation:

(9)

Passives and Syntax

In the literature, several problematic aspects of the above tests have been discussed, and the reader is referred to Alexiadou, Anagnostopoulou, and Schäfer (2015) for a summary of this discussion. For instance, it could very well be that some of the tests simply show that anticausatives like intentional external arguments, as argued for explicitly in, e.g., Kallulli (2007). They could in principle allow causer implicit external arguments in anticausatives. As causers are not intentional external arguments, the above tests have nothing to contribute to the debate concerning the presence vs. absence of implicit external arguments in anticausatives as opposed to passives. What they simply show is that anticausatives lack intentional implicit external arguments. Thus what is needed is a test that is insensitive to ±intentionality/agentivity. Such a test is discussed in Alexiadou, Anagnostopoulou, and Schäfer (2015), namely the use of the adverbial by itself phrase. This test is used in Levin and Rappaport Hovav (1995), building on Chierchia (1989). The observation is the following: by itself is licensed in transitives, and in anticausatives, but crucially not in passives (10). Thus the question arises why this is the case. Now notice that by-itself has two readings: the first one is ‘alone’, the second one, as Alexiadou, Anagnostopoulou, and Schäfer argue, is best characterized as ‘no particular cause’. Only the second reading is relevant here.

(10)

Passives and Syntax

The interpretation of (10b) involves denial of the existence of an external cause that caused the event expressed by the predicate. (10c) means that no entity can be identified to have forced or caused John to break the door. Now consider the passive example in (10a). In this case, the adverbial has the same contribution as in (10b), but now a contradiction arises as the passive contains an implicit external argument that ‘specifies a causer, while by itself denies the presence of a causer’. Thus the authors conclude passives but not anticausatives contain an implicit external argument.

The important observation is that this test, unlike some of the other diagnostics, is not sensitive to agentivity but rather to the presence of causation. As Alexiadou, Anagnostopoulou, and Schäfer (2015) point out, this is particularly clear with verbs that only allow natural forces as external arguments and still undergo the causative alternation. Consider (11), taken from Alexiadou, Anagnostopoulou, and Schäfer (2015). As we see, the anticausative form is legitimate, but the passive form is out:

(11)

Passives and Syntax

Thus we can safely conclude that passives, but not anticausatives, contain an implicit external argument. The question that arises is of course how the implicit external argument is represented in (short) passives. While Collins (2005) argues that this is syntactically projected, Bruening (2012) proposes that this is actually not the case. As Bhatt and Pancheva (2006) and Landau (2010), among others, make clear, this is a very controversial matter. I will not offer a detailed discussion of this controversy here. The reader is referred to Alexiadou, Anagnostopoulou, and Schäfer for discussion on this issue. See also Müller (2015), who shows that the implicit external argument of passives is not accessible by higher heads and is only visible from lower structure.

4 Passive and Voice

4.1 Kratzer

In a seminal paper, Kratzer (1996) proposes that external arguments are not arguments of the verbal predicate, but rather are introduced in Spec,VoiceP. By contrast, direct DP objects are arguments of V and hence are generated within the VP, as in (12). In other words, external arguments are severed from the verbal predicate and are introduced by a functional head, labeled Voice in Kratzer.

(12)

Passives and Syntax

The type of arguments Kratzer brought to support the above syntactic representation can be summarized as follows. To begin with, as discussed also in Marantz (1984), the combination of a verb with a particular type of internal argument seems to be responsible for the thematic role of the predicate’s external argument:

(13)

Passives and Syntax

Second, several researchers have pointed out (Grimshaw, 1990; see also Alexiadou, 2001; Borer, 2013; Siloni, 1997; among others, building on Grimshaw) that certain deverbal nouns must realize the verb’s direct object. This is the case only as long the original eventive interpretation of the verb is preserved, as in (14).

(14)

Passives and Syntax

Grimshaw’s observation is that the presence of the internal argument is obligatory in the above context, which triggers the eventive interpretation for the nominalization, evidenced by the presence of the modifier constant. From these two, and other environments, Kratzer concludes that while a verb has an internal argument as part of its meaning, external arguments are not part of the meaning of the predicate and hence should be severed. Kratzer proposes that they are introduced by the functional head Voice; see Borer (2005) and Lohndal (2014) for arguments in favor of complete thematic separation.

4.2 PassiveP

Kratzer’s intuition is that the same head that introduces the overt external argument in the active is responsible for the introduction of the implicit external argument in the passive (see also Baker & Vinokurova, 2009; Legate, 2014). Schäfer (2008) offers an analysis of transitivity alternations on the basis of different features on this Voice head, namely D and thematic features. Their distribution derives passives and anticausatives across languages.

This analysis makes a clear prediction: if it is indeed the case that active and passive Voice are two versions of the same head, then they should both introduce the same sets of thematic roles. In English, this is certainly the case, as shown in (15) and (15’), from Marantz (1984, p. 129):

(15)

Passives and Syntax

(15’)

Passives and Syntax

However, this is not the case in other languages such as Hebrew and Greek; see (16), from Doron (2015):

(16)

Passives and Syntax

This is one type of argument that was brought against the view that passive and active Voice are two versions of the same head. Researchers who consider Passive to involve a designated head different from Voice include Doron (2003), Collins (2005), Bruening (2012), Alexiadou and Doron (2012), Kiparsky (2013), Merchant (2013), and Alexiadou, Anagnostopoulou, and Schäfer (2015), among others, though the perspectives differ. Some of these analyses take passives to involve a functional head on top of the functional projection Voice that introduces the external argument of the predicate. Specifically, Doron (2003), Alexiadou and Doron (2012), and Alexiadou, Anagnostopoulou, and Schäfer (2015) propose that there are two functional heads that correspond to Kratzer’s nonactive variant of Voice, and their distribution varies cross-linguistically. PassiveP is present in, e.g., English, and some Semitic languages, but nonactive Voice, labeled Middle in these works, is present in languages such as Greek and some other Semitic languages.

Collins (2005), as well as Bruening (2012), also put forth the idea that the head responsible for the formation of passives is above the layer introducing the external argument. Collins labels this head Voice, while Bruening labels it Passive. However, these authors do not discuss a possible parametrization of this head, i.e., the possibility that it can be present in some but not all languages. Gehrke and Grillo (2009) share some similarities with Collins (2005); they differ in that they pay particular attention to the aspectual properties of the passive. From their perspective, passive formation involves the availability of a consequent state reading, which then would explain why some predicates do not form good passives. It is not clear how this approach would deal with the data discussed in the next sub-section.

4.3 The Cross-linguistic Distribution of the Two Syntactic Structures for Passives

An important insight of Kratzer’s (1996) and Embick’s (1998) view of the passive is that active and passive are not in a dependency relation. By contrast, from the viewpoint of proposals that view VoiceP as being embedded under Passive, at least in some languages, the insight is very different: passive formation in languages that have this particular head is based on the availability of a transitive input, i.e., a substructure that contains an external argument; see the discussion of Chomsky (1957) in Section 2. Thus the availability of PassiveP is parametrized; it is present in some languages but not in others.

Several types of argument were discussed in support of this latter view. First, languages that have a passive head differ in terms of the productivity of their passives. For instance, in English nearly all transitive verbs can form a passive variant; by contrast, in languages such as Greek and Semitic, passivization is restricted to some but not all verbs. Second, there is variation concerning the thematic role of the external argument. As we saw above, in languages such as Hebrew the thematic role of the by-phrase is restricted. Similar observations hold for Greek. By contrast, in, e.g., English, this can bear the thematic role of the external argument of the active. In other words, it can be an agent, a causer or an experiencer. But in Greek or Hebrew it can only be an agent, even if the active allows a variety of thematic roles. Third, in some languages not all types of noun phrases can be introduced via a by-phrase. As we will see below, in e.g., Greek by-phrases are more acceptable, if the DP in the by-phrase is nonspecific.

These properties correlate with a formal one. In languages of the English type, call them group I languages, where we have productive passivization, their passives can only be used as passives. In languages of the Greek type, call them group II languages, the Voice morphology used to build passives also surfaces on the formation of other alternations and is even used in forms that are not part of an alternation. On the basis of these observations, we can conclude that there can be no singly mechanism to form passives across languages.

Let me illustrate these differences in some detail on the basis of a comparison between English, Greek, and Semitic (two dialects of Arabic and Hebrew). Laks (2009) provided a detailed description of passive formation in two dialects of Arabic, namely Modern Standard Arabic and Palestinian Arabic. In the former dialect, as in English, every transitive verb can passivize. In this particular dialect passive formation works as follows: the vocalic pattern of a transitive verb changes into u-i and u-a in the perfective and imperfective verbs, respectively, as shown in (17).

(17)

Passives and Syntax

The important observation Laks makes is that passive formation is exception-free and passive verbs have an exclusive passive interpretation. No other type of alternation can be introduced via these vocalic patterns. Note that this is not exactly the situation in English as out of context the be + participle string is ambiguous between a verbal and an adjectival passive. But this is exactly what we observe in German, where the adjectival passive is morphologically distinct from the verbal passive. Verbal passives which are formed on the basis of werden ‘become’ + participle are exception-free and have a passive meaning only:

(18)

Passives and Syntax

As in MSA, in German the passive construction in (18) cannot host any other type of predicate, i.e., it cannot yield a reflexive or an anticausative construal.

By contrast, according to Laks (2009), in Palestinian Arabic (PA), passivization is not very productive. In fact it is possible only in two verbal templates, the so-called facal, and faccal. Unlike MSA, passive formation in these two templates involves adding the prefix in- or t- to the active verb, respectively, as shown in (19):

(19)

Passives and Syntax

But even within these two templates, there are some transitive verbs that simply do not have passive counterparts. In some cases, this seems to be so for no apparent reason:

(20)

Passives and Syntax

The form that is used to form the passive can have a variety of other meanings. Laks notes that the tfaccal template can be used in reflexive and anticausative construals, while the infacal template is primarily used for passive and anticausative predicates. Both templates can host basic deponent verbs, i.e., predicates that do not have a transitive variant:

(21)

Passives and Syntax

In some cases, Laks observes, one can attribute the lack of a passive to morpho-phonology. For instance, if the form would be too complex, it is then blocked.

Laks’s analysis is cast within the Reinhart and Siloni (2005) framework and can be summarized as follows: passive verbs in PA are derived directly from their active variants in the lexicon by applying word formation rules on existing words, when the application is possible. When the application is impossible, however, no passive verbal form is built. In contrast, passives in MSA are built in the syntax and every transitive verb can have a passive counterpart. This naturally makes sense from the perspective of the lexicon–syntax divide. Lack of transparency, the low productivity and irregularity associated with PA-type passives, would suggest that their thematic operations are lexical. What is parametrized from this perspective is the locus of passive derivations: these are lexical in some languages but syntactic in others.

A further language where a similar state of affairs is observed is Greek; see Tsimpli (1989), Embick (1998), Zombolou (2004), and Alexiadou, Anagnostopoulou, and Schäfer (2015) for discussion and references. In Greek, passive formation is, as already mentioned, synthetic as in Arabic. The language has two Voice paradigms, namely Active and Non-Active Voice, (22). Passive verbs are built on the basis of non-active Voice, (23b).

(22)

Passives and Syntax

(23)

Passives and Syntax

As was the case in PA, in Greek nonactive morphology is used in a variety of environments: it is found with certain anticausatives, dispositional middles, all reflexives, and deponent verbs; see Tsimpli (1989), Embick (1998), Alexiadou and Anagnostopoulou (2004), Zombolou (2004), Alexiadou and Doron (2012), and Alexiadou, Anagnostopoulou, and Schäfer (2015), among others:

(24)

Passives and Syntax

Again, similarly to what has been observed for PA, many Greek verbs do not passivize, although their counterparts in English and German can build perfect passives. Zombolou (2004) and Alexiadou, Anagnostopoulou, and Schäfer (2015) note the following restrictions. First, most change-of-state verbs, but also verbs belonging to other verb classes as well, simply cannot form a passive; see the list in (26):

(26)

Passives and Syntax

Alexiadou, Anagnostopoulou, and Schäfer (2015) point out that, as in PA, in several cases passive formation is out due to morpho-phonological constraints. In other words, with some verbs, the combination of their stem with the nonactive affix leads to a phonological clash, i.e., a too complex internal consonant cluster involving an alveolar nasal and a dental fricative would be formed, and Greek speakers do not seem to tolerate that:

(27)

Passives and Syntax

As already mentioned and as we saw above for Hebrew, by-phrases are severely restricted in Greek, either considered marked or only possible if the DP in the by-phrase is nonspecific; see Philippaki-Warburton (1975), Laskaratou and Philippaki-Warburton (1984), and Zombolou (2004).

(28)

Passives and Syntax

The above facts led several researchers to conclude that passives in Greek are lexical and not syntactic (see, e.g., Smyrniotopoulos, 1992). Other authors, for instance, Klaiman (1991), Kaufmann (2004), and Manney (2000), all argue that Greek has middle Voice and lacks a designated passive Voice. As Kemmer (1993) shows in great detail, Middle Voice subsumes a variety of readings, unlike passive Voice, which has a passive only interpretation.

A more recent version of the middle Voice theory is put forth in Doron (2003), Alexiadou and Doron (2012), Alexiadou, Anagnostopoulou, and Schäfer (2015), and Spathas, Alexiadou, and Schäfer (2015). These authors propose that there are two syntactic nonactive Voice heads involved across languages in the formation of passives. The first head is labeled Passive and always takes a structure, which includes VoiceP as its input. The second head is labeled Middle and is the nonactive counterpart of Kratzer’s Voice head. This type of head is underspecified in the sense of Embick (1998): it always surfaces with nonactive morphology as the only structural requirement passives, anticausatives, reflexives, and dispositional middles share in this language is that they all lack an overt external argument, i.e., they are unaccusative structures (for middles, see Lekakou, 2005; for reflexives, see Alexiadou & Schäfer, 2014).

The discussion thus far suggests that the availability of these two heads is subject to parametric variation. However, Doron (2003) shows that this is not the case. She demonstrates in great detail that Hebrew has both these heads. Based on her insights and observations, the distribution of these two heads can be summarized as follows. In Hebrew, each root can appear in combination with three types of verbal templates: the simple, the intensive, and the causative. Depending on the verbal template a particular root appears in, the external argument of the predicate will receive a different interpretation, but I will not go into this discussion here. Importantly, these agency templates, as Doron calls them, can combine with three Voice templates: active, middle, and passive. Doron observes that there are gaps in these combinations, so the simple template lacks a passive Voice and the causative template a middle one:

(29)

Passives and Syntax

What is relevant for our discussion is the interpretation of these agency-Voice combinations. According to Doron, middle morphology, which is found in the simple and intensive template, can have anticausative and reflexive interpretations. By contrast, in the causative template there is no verb that can have a reflexive or anticausative interpretation in the passive Voice.

(30)

Passives and Syntax

As expected, in view of our earlier discussion, the Hebrew middle template can have a passive interpretation; see (31), from Alexiadou and Doron (2012):

(31)

Passives and Syntax

But in the intensive template, the passive can have a passive only interpretation; see (32), again from Alexiadou and Doron (2012). As the ungrammaticality of the by-itself phrase shows, this particular form cannot have an anticausative interpretation, (32b).

(32)

Passives and Syntax

This seems to provide evidence for the view that Middle and Passive can co-occur in a language. (33) gives a structural illustration of these two structures. From the perspective of (33a), passive is an operation on active transitive verb phrases, while this is not the case in (33b). This structure is underdetermined for the semantic interpretation it can receive: depending on the type of verb included, it can yield a reflexive, a dispositional middle, an anticausative, or a passive interpretation. Importantly, this structure is not dependent on there being a transitive entry, so deponent verbs can be created:

(33)

Passives and Syntax

Recently, Oikonomou (2014) made the following observation that attempts to link the parametrization to the presence of analytic vs. synthetic passives. In languages such as Greek and PA, synthetic passives can receive additional interpretations beyond the prototypical passive, while analytic passives cannot. What this generalization excludes is the possibility of an analytic passive to receive additional interpretations. At first sight, this hypothesis seems to face problems with Hebrew and MSA, which both have synthetic passives, but in which the passive template has a passive only interpretation. In the next section, we will see further problems for this view. As we will see, even English provides evidence that its analytic passive can receive a different set of interpretations: it can be an adjectival as well as a verbal passive. This seems to suggest that the distinction one needs to draw in order to capture the properties of (all types of) passives cross-linguistically does relate to their syntactic building blocks.

5 Verbal vs. Adjectival Passives

In Section 3, we discussed the fact that a basic difference between verbal and adjectival passives relates to the presence of an implicit external argument in the former as opposed to the latter. This basic difference was interpreted by Kratzer (1996) and most prominently by Anagnostopoulou (2003), as suggesting that adjectival passives unlike verbal ones lack Voice, the head that introduces the external argument of the predicate. Earlier literature on the differences between the two constructions suggested by contrast that verbal passives are derived in the syntax, while adjectival passives are derived in the lexicon.

Let us briefly summarize these differences and review some earlier as well as some recent views on adjectival passives. Wasow (1977) pointed out that there are several systematic differences between the two constructions that can lead us to propose that these are generated in different grammar modules. To begin with, adjectival passives may show irregular morphology, while verbal passives always show regular morphology.

(34)

Passives and Syntax

While adjectival passives can have idiomatic meanings, this is not the case for verbal passives:

(35)

Passives and Syntax

Importantly, adjectival passive formation does not interact with syntactic operations. For instance, there is no ECM followed by passivization with adjectival passives, but there is such an operation with verbal passives:

(36)

Passives and Syntax

On the basis of these and other differences, Wasow concluded that adjectival passives are built in the lexicon, while verbal passives are derived in the syntactic component. This explains why the former show idiosyncratic form and meaning, while the latter are compositional and productive: word formation, as we stated already, in the lexicon has special properties, syntactic operations are by contrast completely transparent and regular. For some further discussion on the lexical properties of adjectival passives, see Bresnan (1982), Levin and Rappaport Hovav (1986), and Williams (1981), who all assume a lexical rule for adjectival passive formation.

This picture was revised by Kratzer (1996), Anagnostopoulou (2003), and Embick (2004), who all pointed out that adjectival passives are not uniform within a language and across languages. While the details of these contributions differ, the common intuition is that certain adjectival passives are closer to verbal ones in the sense that contain event implications and are not simply stative. In addition, Anagnostopoulou (2003) argued that the availability of implicit external arguments in adjectival passives is parametrized: adjectival passives in Greek, but not in English and German, provide evidence for the presence of a Voice head.

Perhaps the clearest contrast with respect to the first property is the one provided in Embick (2004), which is discussed in great detail in Alexiadou, Anagnostopoulou, and Schäfer (2015). Consider the following examples. (37a) has two readings: the first one Embick calls the eventive/verbal passive, and the second one he labels the resultative passive. By contrast, (37b) has only the simple stative reading:

(37)

Passives and Syntax

Simple states cannot license manner adverbial modification, while resultatives can:

(38)

Passives and Syntax

Another difference between simple states and resultative forms concerns the fact that former can occur after verbs of creation while the latter cannot:

(39)

Passives and Syntax

The reason for this is that it would be a contradiction if the complement of a verb of creation were a state resulting from a prior event.

While in the case of open, there is a special form for the simple state, for other predicates such as closed the same participial form is used for all three readings. As we see in (40a), closed can appear in the context of a verb of creation. It can also have a resultative interpretation, as can be seen from the fact that it can be modified by manner adverbials, (40b):

(40)

Passives and Syntax

This suggests that closed is three-way ambiguous. Embick (2004) proposed that is because the syntactic composition of these three readings is different. Simple states are root-=derived adjectives, while resultative participles are verb-derived. Eventive passives differ in that they contain a head that introduces the implicit external argument, i.e., Voice in our terminology.

Anagnostopoulou (2003) discussed in detail how Greek adjectival passive formation differs from its counterpart in English and German. Specifically, in Greek a particular type of participle, which Anagnostopoulou (2003), following Kratzer (2000), shows to be resultant state participles, provide evidence of the presence of Voice. For instance, such participles accept agent-oriented modification as well as by-phrases:

(41)

Passives and Syntax

On the basis of examples such as the above, Anagnostopoulou (2003) proposed that Voice can only be present in Greek resultant state participles. By contrast, Voice is absent from German and English resultant state adjectival passives.

In more recent literature, however, this claim has been refuted. For instance, McIntyre (2013) and Bruening (2014) show that adjectival participles in English can include Voice. Alexiadou, Gehrke, and Schäfer (2013) make a similar claim for German. In both languages it can shown on the basis of, e.g., adverbial modification, licensing of by-phrases, and control into purpose clauses that adjectival passives contain Voice. Let us consider these arguments.

McIntyre (2013), building on Meltzer-Asscher (2011), argues that some adjectival participles in English actually have implicit external arguments. McIntyre focuses on the licensing of by-phrases. As he demonstrates, by-phrases are licit in some English adjectival passives, but only if the by-phrase referent is ‘responsible for continuing the state expressed by the participle.’ This is shown in the following examples (from McIntyre (2013), his 19):

(42)

Passives and Syntax

A similar point has been made for German; consider (43), from Maienborn (2007), and see references therein as well as in Alexiadou, Anagnostopoulou, and Schäfer (2015). Since German distinguishes morphologically between verbal and adjectival passives, examples such as (43) suggest that implicit external arguments cannot be taken to distinguish between the two, as they can be present in both. For a more detailed discussion on these and other contrasts, see Alexiadou, Anagnostopoulou, and Schäfer (2015, chapter 5).

(43)

Passives and Syntax

If this is indeed the case, the question that arises is whether there is any structural difference between adjectival and verbal passives. As we saw above, however, a major difference between the two relates to the presence of disjoint reference effects in verbal passives, and the absence thereof in adjectival ones. How can we then reconcile these two contradictory observations? Similar concerns arise with respect to the English get-passive, and the so-called recipient passive across languages, which I will not discuss here. See Alexiadou (2012), Alexiadou, Anagnostopoulou, and Sevdali (2014), and the contributions to the special issue on get verbs in Linguistics 50 on recipient passives; see the contributions to Alexiadou and Schäfer (2013) on other types of noncanonical passives.

A way to approach this issue in the context of adjectival passives is offered in Alexiadou, Anagnostopoulou, and Schäfer (2015). These authors point out that what we see in the context of adjectival passives is precisely the situation that we have in Middle as opposed to Passive Voice languages. In Middle Voice type languages, the presence vs. absence of a disjoint reference effect is simply not encoded on the particular head involved in the syntactic construction, but must be resolved by local context, e.g., the semantic properties of the root or verb involved. Thus, it could very well be that adjectival passives instantiate actually Middle Voice structures, thus providing further evidence against Oikonomou’s analysis proposed above.

Further Reading

Åfarli, T. (1992). The syntax of Norwegian passive constructions. Amsterdam: John Benjamins.Find this resource:

Alexiadou, A., Anagnostopoulou, E., & Schäfer, F. (2015). External arguments in transitivity alternations: A layering approach. Oxford: Oxford University Press.Find this resource:

Alexiadou, A., & Schäfer, F. (Eds.). (2013). Non-canonical passives. Amsterdam: John Benjamins.Find this resource:

Baker, M., Johnson, K., & Roberts, I. (1989). Passive arguments raised. Linguistic Inquiry, 20, 219–251.Find this resource:

Bhatt, R., & Pancheva, R. (2006). Implicit arguments. In M. Everaert & H. van Riemsdijk (Eds.), The Blackwell companion to syntax (pp. 554–584). Malden, MA: Blackwell.Find this resource:

Bresnan, J. (1982). The passive in lexical theory. In J. Bresnan (Ed.), The mental representation of grammatical relations (pp. 3–86). Cambridge, MA: MIT Press.Find this resource:

Bruening, B. (2012). By-phrases in passives and nominals. Syntax, 16, 1–41.Find this resource:

Collins, C. (2005). A smuggling approach to the passive in English. Syntax, 8(2), 81–120.Find this resource:

Doron, E. (2003). Agency and voice: The semantics of semitic templates. Natural Language Semantics, 11, 1–67.Find this resource:

Doron, E. (2015). Voice and valence change. In T. Kiss & A. Alexiadou (Eds.), Syntax: Empirical and theoretical approaches (pp. 749–776). Berlin: Walter de Gruyter.Find this resource:

Gehrke, B., & Grillo, N. (2009). How to become passive. In K. Grohmann (Ed.), Explorations of phase theory: Features, arguments, and interpretation at the interfaces (pp. 231–268). Berlin: de Gruyter.Find this resource:

Huang, J. C-T. (1999). Chinese passive in comparative perspective. Tsing Hua Journal of Chinese Studies, 29, 423–509.Find this resource:

Keenan, E., & Dryer, M. (2006). Passive in the world’s languages. In T. Shopen (Ed.), Clause structure, language typology and syntactic description (2d ed., Vol. 1, pp. 325–361). Cambridge, U.K.: Cambridge University Press.Find this resource:

Kiparsky, P. (2013). Towards a null theory of the passive. Lingua, 125, 7–33.Find this resource:

Kratzer, A. (1996). Severing the external argument from its verb. In J. Rooryck & L. Zaring (Eds.), Phrase structure and the lexicon (pp. 109–137). Dordrecht, The Netherlands: Kluwer.Find this resource:

Kratzer, A. (2000). Building statives. Proceedings of BLS, 26, 385–399.Find this resource:

Lyngfelt, B., & Solstad, T. (Eds.). (2006). Demoting the agent: Passive, middle and other voice phenomena. Amsterdam: John Benjamins.Find this resource:

McIntyre, A. (2013). Adjectival passives and adjectival participles in English. In A. Alexiadou & F. Schäfer (Eds.), Non-canonical passives (pp. 21–42). Amsterdam: John Benjamins.Find this resource:

References

Åfarli, T. (1992). The syntax of Norwegian passive constructions. Amsterdam: John Benjamins.Find this resource:

Alexiadou, A. (2001). Functional structure in nominals: Nominalization and ergativity. Amsterdam: John Benjamins.Find this resource:

Alexiadou, A. (2012). Non-canonical passives revisited: Parameters of non-active voice. Linguistics, 50, 1079–1110.Find this resource:

Alexiadou, A. & Anagnostopoulou, E. (2004). Voice morphology in the causative-inchoative Alternation: evidence for a non unified structural analysis of unaccusatives. In A. Alexiadou, E. Anagnostopoulou, & M. Everaert (Eds.), The unaccusativity puzzle (pp. 22–59). Oxford: Oxford University Press.Find this resource:

Alexiadou, A., Anagnostopoulou, E., & Schäfer, F. (2015). External arguments in transitivity alternations: A layering approach. Oxford: Oxford University Press.Find this resource:

Alexiadou, A., Anagnostopoulou, E., & Sevdali, C. (2014). Opaque and transparent datives, and how they behave in passives. The Journal of Comparative Germanic Linguistics, 17, 1–34.Find this resource:

Alexiadou, A., & Doron, E. (2012). The syntactic construction of two non-active voices: Passive and middle. Journal of Linguistics, 48, 1–34.Find this resource:

Alexiadou, A., Gehrke, B., & Schäfer, F. (2013). The argument structure of adjectival participles revisited. Lingua, 149, 118–138.Find this resource:

Alexiadou, A., & Schäfer, F. (Eds.). (2013). Non-canonical passives. Amsterdam: John Benjamins.Find this resource:

Alexiadou, A., & Schäfer, F. (2014). Towards a non-uniform analysis of naturally reflexive verbs. Proceedings of WCCFL, 31, 1–10.Find this resource:

Anagnostopoulou, E. (2003). Participles and voice. In A. Alexiadou, M. Rathert, & A. von Stechow (Eds.), Perfect explorations (pp. 1–36). Berlin: Mouton de Gruyter.Find this resource:

Baker, M., Johnson, K., & Roberts, I. (1989). Passive arguments raised. Linguistic Inquiry, 20, 219–251.Find this resource:

Baker, M., & Vinokurova, N. (2009). On agent nominalizations and why they are not like event nominalizations. Language, 85, 517–555.Find this resource:

Bhatt, R., & Pancheva, R. (2006). Implicit arguments. In M. Everaert & H. van Riemsdijk (Eds.), The Blackwell companion to syntax (pp. 554–584). Malden, MA: Blackwell.Find this resource:

Borer, H. (2005). The normal course of events. Oxford: Oxford University Press.Find this resource:

Borer, H. (2013). Taking form. Oxford: Oxford University Press.Find this resource:

Bresnan, J. (1982). The passive in lexical theory. In J. Bresnan (Ed.), The mental representation of grammatical relations (pp. 3–86). Cambridge, MA: MIT Press.Find this resource:

Bruening, B. (2012). By-phrases in passives and nominals. Syntax, 16, 1–41.Find this resource:

Bruening, B. (2014). Word formation is syntactic: Adjectival passives in English. Natural Language and Linguistic Theory, 32, 363–422.Find this resource:

Burzio, L. (1986). Italian syntax: A government and binding approach. Dordrecht, The Netherlands: Reidel.Find this resource:

Chierchia, G. (1989/2004). A semantics for unaccusatives and its syntactic consequences. In A. Alexiadou, E. Anagnostopoulou, & M. Everaert (Eds.), The unaccusativity puzzle: Explorations of the syntax-lexicon interface (pp. 22–59). Oxford: Oxford University Press.Find this resource:

Chomsky, N. (1957). Syntactic structures. Berlin: Mouton de Gruyter.Find this resource:

Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (1981). Lectures on government and binding. Dordrecht, The Netherlands: Foris.Find this resource:

Collins, C. (2005). A smuggling approach to the passive in English. Syntax, 8(2), 81–120.Find this resource:

Doron, E. (2003). Agency and voice: The semantics of semitic templates. Natural Language Semantics, 11, 1–67.Find this resource:

Doron, E. (2015). Voice and valence change. In T. Kiss & A. Alexiadou (Eds.), Syntax: Empirical and theoretical approaches (pp. 749–776). Berlin: Walter de Gruyter.Find this resource:

Embick, D. (1998). Voice systems and the syntax/morphology interface. Papers from the UPenn/MIT roundtable on argument structure and aspect, 32, 41–72.Find this resource:

Embick, D. (2004). On the structure of resultative participles in English. Linguistic Inquiry, 35, 355–392.Find this resource:

Eythórsson, T. (2008). The new passive in Icelandic is really a passive. In T. Eythórsson (Ed.), Grammatical change and linguistic theory: The Rosendal papers (pp. 173–219). Amsterdam: John Benjamins.Find this resource:

Gehrke, B., & Grillo, N. (2009). How to become passive. In K. Grohmann (Ed.), Explorations of phase theory: Features, arguments, and interpretation at the interfaces (pp. 231–268). Berlin & New York: de Gruyter.Find this resource:

Grimshaw, J. (1990). Argument structure. Cambridge, MA: MIT Press.Find this resource:

Haspelmath, M. (1990). The grammaticization of passive morphology. Studies in Language, 14(1), 25–71.Find this resource:

Kallulli, D. (2007). Rethinking the passive/anticausative distinction. Linguistic Inquiry, 38(4), 770–780.Find this resource:

Kaufmann, I. (2004). Medium und Reflexiv—Eine Studie zur Verbsemantik. Tübingen, Germany: Niemeyer.Find this resource:

Keenan, E., & Dryer, M. (2006). Passive in the world’s languages. In T. Shopen (Ed.), Clause structure, language typology and syntactic description (2nd ed., Vol. 1, pp. 325–361. Cambridge, U.K.: Cambridge University Press.Find this resource:

Kemmer, S. (1993). The middle voice. Amsterdam: John Benjamins.Find this resource:

Kiparsky, P. (2013). Towards a null theory of the passive. Lingua, 125, 7–33.Find this resource:

Klaiman, M. H. (1991). Grammatical voice. Cambridge, U.K.: Cambridge University Press.Find this resource:

Kratzer, A. (1996). Severing the external argument from its verb. In J. Rooryck & L. Zaring (Eds.), Phrase structure and the lexicon (pp. 109–137). Dordrecht, The Netherlands: Kluwer.Find this resource:

Kratzer, A. (2000). Building statives. Proceedings of BLS, 26, 385–399.Find this resource:

Laks, L. (2009). The formation of Arabic passive verbs: Lexical or syntactic? Proceedings of IATL, 25.Find this resource:

Landau, I. (2010). The explicit syntax of implicit arguments. Linguistic Inquiry, 41, 357–388.Find this resource:

Laskaratou, C., & I. Philippaki-Warburton (1984). Lexical vs. transformational passives in Modern Greek. Glossologia, 2–3, 99–109.Find this resource:

Lavine, J. E. (2013). Passives and near passives in Balto-Slavic: On the survival of accusatives. In A. Alexiadou & F. Schäfer (Eds.), Non-canonical passives (pp. 185–211). Amsterdam: John Benjamins.Find this resource:

Legate, J. (2014). Voice and v: Lessons from Acehnese. Cambridge, MA: MIT Press.Find this resource:

Lekakou, M. (2005). In the middle, somewhat elevated. The semantics of middles and its crosslinguistic realization (Unpublished PhD Diss.). University of London.Find this resource:

Levin, B., & Rappaport Hovav, M. (1986). The formation of adjectival passives. Linguistic Inquiry, 17, 623–661.Find this resource:

Levin, B., & Rappaport Hovav, M. (1995). Unaccusativity: at the syntax-lexical semantics interface. Cambridge, MA: MIT Press.Find this resource:

Lohndal, T. (2014). Phrase structure and argument structure: A case study of the syntax-semantics interface. Oxford: Oxford University Press.Find this resource:

Maienborn, C. (2007). Das Zustandspassiv: Grammatische Einordnung—Bildungsbeschränkungen—Interpretationsspielraum. Zeitschrift für Germanistische Linguistik, 35, 83–115.Find this resource:

Maling, J., & Sigurjónsdóttir, S. (2002). The ‘new impersonal’ construction in Icelandic. Journal of Comparative Germanic Linguistics, 5, 97–142.Find this resource:

Manney, L. J. (2000). Middle voice in modern Greek: Meaning and function of an inflectional category. Amsterdam: John Benjamins.Find this resource:

Marantz, A. (1984). On the nature of grammatical relations. Cambridge, MA: MIT Press.Find this resource:

McIntyre, A. (2013). Adjectival passives and adjectival participles in English. In A. Alexiadou & F. Schäfer (Eds.), Non-canonical passives (pp. 21–42). Amsterdam: John Benjamins.Find this resource:

Meltzer-Asscher, A. (2011). Adjectival passives in Hebrew: Evidence for parallelism between the adjectival and verbal systems. Natural Language and Linguistic Theory, 29, 815–855.Find this resource:

Merchant, J. (2013). Voice and ellipsis. Linguistic Inquiry, 44, 77–108.Find this resource:

Müller, G. (2015). Structure removal: a new approach to conflicting representations.Find this resource:

Müller, S. (2007). Head-driven phrase structure grammar: Eine Einführung. Tübingen, Germany: Stauffenberg.Find this resource:

Oikonomou, D. (2014). In the middle of passive: Middle voice in Modern Greek vs. passive voice in English. Ms., MIT.Find this resource:

Perlmutter, D., & Postal, P. (1983). Towards a universal characterization of passivization. In D. Perlmutter (Ed.), Studies in relational grammar (pp. 3–29). Chicago: University of Chicago Press.Find this resource:

Philippaki-Warburton, I. (1975). The passive in English and Greek. Foundations of Language, 13, 563–578.Find this resource:

Pitteroff, M. (2015). Morphosyntactically unmarked passives: Evidence from let-passives and let-middles. Proceedings of NELS, 45.Find this resource:

Reinhart, T., & Siloni, T. (2005). The lexicon-syntax parameter: Reflexivization and other Arity operations. Linguistic Inquiry, 36, 389–436.Find this resource:

Schäfer, F. (2008). The syntax of (anti-)causatives: External arguments in change-of-state contexts. Amsterdam, The Netherlands: John Benjamins.Find this resource:

Siloni, T. (1997). Noun phrases and nominalization. Dordrecht, The Netherlands: Foris.Find this resource:

Smyrniotopoulos, J. (1992). Lexical passives in modern Greek. Garland.Find this resource:

Sobin, N. (1985). Case assignment in Ukrainian morphological passive constructions. Linguistic Inquiry, 16, 649–662.Find this resource:

Spathas, G., Alexiadou, A., & Schäfer, F. (2015). Middle voice and reflexive interpretations: Afto-prefixation in Greek. Natural Language and Linguistic Theory.Find this resource:

Tsimpli, I. M. (1989). On the properties of the passive affix in modern Greek. UCL Working Papers in Linguistics, 1, 235–260.Find this resource:

Wasow, T. (1977). Transformations and the Lexicon. In P. W. Culicover, T. Wasow, & J. Bresnan (Eds.), Formal syntax (pp. 327–360). New York: Academic Press.Find this resource:

Williams, E. (1981). Argument structure and morphology. The Linguistic Review, 1, 81–114.Find this resource:

Wunderlich, D. (1993). Diathesen. In J. Jacobs, A. von Stechow, W. Sternefeld, & T. Vennemann (Eds.), Syntax. Ein Internationales Handbuch zeitgenössischer Forschung I (pp. 730–747). Berlin: Mouton de Gruyter.Find this resource:

Zombolou, K. (2004). Verbal alternations in Greek: A semantic analysis (Unpublished PhD Diss.). University of Reading, Reading, United Kingdom.Find this resource: