1 Introduction

Since Kratzer (1996)’s influential proposal that the external argument is truly external as it is introduced in the syntax independently of the verb, by means of a functional head that Kratzer labelled Voice, a considerable amount of distinct approaches have pursued the idea that a more fine-grained classification is needed regarding the nature of this functional head (cf., D’Alessandro et al. 2017). Evidence in favor of this view mainly comes from observing that there appear to be distinct types of external arguments which exhibit distinct syntactic and semantic properties.

In this paper, we attempt to provide additional novel evidence in favor of such a view by observing that the subjects of some Spanish and Catalan reflexive verbs like matarse ‘kill’, electrocutarse ‘electrocute’ or herirse ‘hurt’, when appear in an intransitive construction, have both internal and external argument properties. The data we concern ourselves with comes from Spanish and Catalan and involves examples like Juan se mató en un accidente de coche ‘Juan got himself killed in a car accident’ in which the subject’s referent appears to behave as both an agent and a patient since it needs to be engaged in an action, in this case Juan’s driving, that in turn leads to its accidental death. In order to account for the subject’s hybrid nature, we follow Pineda & Berro (2020) in proposing that mixed thematic interpretations are the result of subjects being assigned more than one semantic role by virtue of being associated with more than one distinct head in the syntax. More specifically, we argue that, in the intransitive construction, the subjects of these verbs initially merge in the specifier position of a small clause subpredicate (Hoekstra 1992), complement of v, and are co-indexed with an unpronounced specifier of a thematic, but syntactically intransitive Voice head (Schäfer 2008; Alexiadou et al. 2015), lexicalized in turn by the pronoun se in Spanish or es in Catalan, which we gloss as se for descriptive ease. The fact that the subjects have both internal and external argument properties is therefore accounted for since Voice assigns its agent argument to the referent of the internal argument, i.e., the subject, as it is co-indexed with the unpronounced specifier of Voice (Pineda & Berro 2020). We note our proposal is capable of accounting for the subject’s internal and external argument properties as the co-indexation makes it possible for the subject to bear an agent and patient semantic role since it is an argument of both Voice and the small clause predicate. Essentially, the present paper provides a replication of the existing proposal for a class of verbs in Basque laid out in Pineda & Berro (see next section for details). We show that Pineda & Berro’s existing account is viable for a class of reflexive verbs in Spanish and Catalan, and therefore provide additional independent evidence from Romance that supports their proposal.

In order to provide evidence in favor of the present analysis, we further consider transitive reflexive uses of the same verbs, e.g., kill oneself, which share the same surface form as the intransitive uses, e.g., El sospechoso del homicidio se mató al estar rodeado por la policía ‘The suspect killed himself when he was surrounded by the police’ (Web). In this respect, we propose that the intransitive and reflexive uses of verbs like matarse in Spanish and Catalan involve a distinct semantic and syntactic configuration, despite involving identical surface form. Main evidence comes from considering a battery of semantic and syntactic diagnostics that probe for the presence of external and internal arguments. These diagnostics show that the subject of these verbs in the intransitive construction behaves as both an external and internal argument, whereas the subject of transitive reflexive uses involves a canonical external argument. For instance, we observe that only intransitive uses entail accidental death, and therefore lack of intentionality, are naturally compatible with bare plural postverbal subjects and allow en cliticization with a patient interpretation of the subject in Romance languages that have cliticization of this sort such as Catalan (Solà 1973; Burzio 1986; Rigau 1990; Levin & Rappaport Hovav 1995). In short, we note that whereas reflexive uses of a verb like matarse generate a suicide interpretation, the intransitive uses generate an interpretation that the subject’s referent does something that ends up killing them, without the intention of actually committing suicide. The data from Spanish and Catalan ultimately allows us to raise the point that agents and external arguments do not always correlate since some verb classes can have identical surface form, despite involving underlyingly distinct argument alignment.

We proceed as follows. In Section 2, we provide a general overview of Kratzer’s (1996) proposal that the external argument is to be severed from the argument structure of the verb. We further discuss the recent view that pushes Kratzer’s agenda even further by arguing that external arguments do not show uniform behavior regarding how they are introduced syntactically. In Section 3, we first provide a description of the grammatical properties of the Spanish and Catalan verbs that appear in the intransitive construction at stake. This allows us in turn to make clear predictions regarding what sort of reflexive verbs in Spanish and Catalan may exhibit this use. We then move on to propose that the intransitive construction has distinct syntactic and semantic properties than transitive reflexives, despite both sharing the same surface form. In Section 4, following Pineda & Berro (2020), we lay out the present formal analyses of both the intransitive and transitive reflexives uses of verbs like matarse in Spanish and Catalan. Section 5 concludes the paper.

2 Introducing the external argument

Theories of argument structure assuming a syntactic decompositional view of verb meaning have long defended the idea that external arguments are not to be considered true arguments of the verbs themselves as they are claimed to be introduced independently of the verbs by functional heads in the syntax such as Voice or little v (Chomsky 1995; Kratzer 1996; Embick 2004; Folli & Harley 2008; Pylkkännen 2008; Ramchand 2008; Harley 2013; Merchant 2013; Alexiadou 2014; Alexiadou et al. 2015; Harley 2017; Schäfer 2017; Sundaresan & McFadden 2017, among many others. Further see D’Alessandro et al. 2017 for a general overview). This assumption is based mostly on Marantz’s (1984) original observation that verbs only appear to impose semantic restrictions on their internal argument, and never on their external argument. Marantz (1984: 25) illustrated this by considering idiomatic meaning of verbs, which are only triggered by internal arguments (1), and never by external arguments (2).

    1. (1)
    1. a.
    1. kill a cockroach.
    1.  
    1. b.
    1. kill a conversation.
    1.  
    1. c.
    1. kill an evening watching TV.
    1.  
    1. d.
    1. kill a bottle (i.e., empty it).
    1.  
    1. e.
    1. kill an audience (i.e., wow them).
    1. (2)
    1. a.
    1. Harry killed DP.
    1.  
    1. b.
    1. Everyone is always killing DP.
    1.  
    1. c.
    1. The drunk refused to kill DP.
    1.  
    1. d.
    1. Silence can certainly kill DP.
    1.  
    1. e.
    1. Cars kill DP.

Marantz observed that whereas the type of internal argument can alter the meaning of the verb, external arguments never do (but see Nunberg et al. 1994; Wechsler 2005 for criticisms on this view). For instance, when kill has the meaning of spend time doing x as in The boy killed the afternoon watching TV, Marantz notes that kill selects an internal argument that must denote time intervals. Drawing on Marantz’s observation, what are generally known as configurational approaches to argument structure (e.g., Borer 2003; Folli & Harley 2005; Alexiadou et al. 2015) have claimed that external arguments are not true arguments of the verbs themselves, but are instead introduced independently of the verb by functional projections in the syntax. Most notably, Kratzer (1996) influentially proposed that external arguments are introduced by the functional head Voice, whereas internal arguments are true arguments of the verbs themselves as they are generated in the specifier position of the VP. On this view, external arguments are truly external to the verb as it is the Voice head which provides an agent role to the external argument that it introduces.

    1. (3)
    1. Mittie fed the dog.                                                  (adapted from Alexiadou et al., 2015: 7)

Since Kratzer’s influential proposal, a number of distinct approaches have defended the idea that external arguments are not introduced and interpreted uniformly. Instead, classes of external arguments, with distinct syntactic and semantic properties, can be differentiated depending on the position in the syntactic structure they are generated in (cf., D’Alessandro et al. 2017). For instance, Alexiadou (2014) has argued in favor of such a view by examining so-called internally caused change-of-state verbs in English, e.g., blossom, ferment, flower (Levin & Rappaport Hovav 1995; Rappaport Hovav 2020). Alexiadou proposes that these verbs do not form a uniform class since two general classes can be differentiated depending on the properties of their subjects. Namely, what Alexiadou calls the ferment class involves verbs whose subject is introduced by Voice, whereas the subject of what Alexiadou calls the blossom class is introduced in vP. Alexiadou notes that this difference in subject types depending on where they are generated is empirically motivated, e.g., the blossom class does not allow passivization, whereas the ferment class does, and only the subject of the blossom class is thematically restricted to the semantic role of causer.

Similarly, Tollan (2018) and Tollan & Oxford (2018) have proposed that the subjects of some verb classes in Polynesian languages and Algonquian are to be introduced by distinct heads as well. More specifically, along the lines of what Alexiadou (2014) proposed for internally caused change-of-state verbs in English, these authors propose that the subjects of unergatives and transitives in these languages are generated in distinct positions. In the case of unergatives, the subjects are generated in vP, while the subjects of transitives are introduced by Voice. Tollan and Tollan and Oxford provide evidence in favor of this view by considering how case marking works along with the fact that the subjects of unergatives and transitives in these languages appear to show distinct semantic properties.

Recently, Pineda & Berro (2020) have analyzed a group of agentive process verbs in Eastern dialects of the Basque language, e.g., bazkaldu ‘have lunch’, solastatu ‘talk’, borrokatu ‘fight’ or jolastu ‘play’ and observed that they show mixed properties related to unergativity and unaccusativity. For instance, while verbs of this class typically behave as unergative (Levin 1983), the group of verbs analyzed by Pineda & Berro nonetheless select an absolutive subject and the auxiliary be, the typical auxiliary verb selection for unaccusatives in languages that show a split auxiliary system for the perfect (cf., Sorace 2000). Pineda & Berro further note that this group of agentive verbs behave like unaccusatives in other respects since they do not allow cognate objects or partitive subjects. In order to account for their mixed properties regarding their status as unergative or unaccusative, Pineda & Berro propose that the subject of this class of agentive verbs is generated within the vP and subsequently co-indexed with a thematic but non-pronounced argument introduced as the specifier of Voice (Schäfer 2008; Alexiadou et al. 2015). Pineda & Berro note that this explains why their subject has both external and internal argument properties, since the co-indexation makes it possible for the subject to be assigned two distinct semantic roles.

In what follows, we first provide a detailed description of the characteristics of the reflexive verbs in Spanish and Catalan that participate in the intransitive construction, which we show form a natural class. As we discuss, this allows us to make predictions about what sort of reflexive verbs in Spanish an Catalan are expected to participate in this construction.

3 Hybrid subjects in Spanish and Catalan

As briefly mentioned above, this paper is concerned with an intransitive use of some reflexive verbs in Spanish and Catalan which involves subjects that show hybrid behavior between agent and patient.1 Intuitively, the subjects’ referents need to be engaged in an action that accidentally leads to the change that they undergo, e.g., Haider’s drunk driving ends up killing him in (4-a). Similar comments apply to the other example.

    1. (4)
    1. a.
    1. Haider
    2. Haider
    1. conducía
    2. drive.3sg.ipfv
    1. ebrio
    2. drunk
    1. cuando
    2. when
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche.
    2. car.
    1. “Haider got himself killed as he was driving under the influence.” (Web)
    1.  
    1. b.
    1. El
    2. the
    1. joven,
    2. youngster
    1. de
    2. of
    1. 15
    2. 15
    1. años,
    2. years
    1. se
    2. se
    1. electrocutó
    2. electrocute.3sg.pfv
    1. cuando
    2. when
    1. […]
    2. […]
    1. intentaba
    2. try.3sg.ipfv
    1. subirse
    2. climb=se
    1. al
    2. at.the
    1. techo
    2. roof
    1. de
    2. of
    1. un
    2. a
    1. tren.
    2. train
    1. “The fifteen-year-old youngster got himself electrocuted when he was trying to get on the roof a train.” (Web)

The Spanish and Catalan reflexive verbs that appear in this intransitive construction all share one key property. Namely, they must encode a result state that comes about accidentally as a result of the action the referent of the subject is engaged in. For instance, in the example above in (4-a), the reflexive verb matarse encodes the result state of being dead which comes about accidentally as an unexpected consequence from the driving action the referent of the subject is currently engaged in.

We understand the notion result state in the sense of Rappaport Hovav & Levin (2010) (see also Rappaport Hovav & Levin 1998; Beavers 2011; Rappaport Hovav 2014; Beavers & Koontz-Garboden 2020). On this view, a result state involves a change in some attribute or property of a patient after the event described by the verb is over. For instance, the result verb die describes the property of being dead which necessarily holds of a patient after they participate in a dying event. Put differently, a result state necessarily involves an altered degree of some property or value of a patient at the end of the event (see Rappaport Hovav & Levin 2010; Rappaport Hovav 2014). In this respect, one diagnostic that probes for result states involves the something is different about x diagnostic laid out by Beavers (2011). The logic behind this diagnostic relates to the idea that if a verb encodes a result state, and hence a change in some value or property, then denying that nothing is different about the patient should generate clear contradictions. This is illustrated below for the result verbs die, shatter and kill.

    1. (5)
    1. a.
    1. #Kim just died, but nothing is different about her.
    1.  
    1. b.
    1. #Kim just shattered the vase, but nothing is different about it.
    1.  
    1. c.
    1. #Kim just killed the prisoner, but nothing is different about him.

Similarly, another way to probe for result states involves explicitly denying the specific result state that is purportedly encoded by the verb (Rappaport Hovav & Levin 2010). For instance, intuitively, the verb die necessarily results in the death of a participant. Hence, denying that the patient does not die after a dying event results in a clear contradiction. Similar comments apply to the other verbs that encode result states involving distinct attributes.

    1. (6)
    1. a.
    1. #Kim just died, but she is not dead.
    1.  
    1. b.
    1. #Kim just cleaned the table, but it is not clean.
    1.  
    1. c.
    1. #Kim just destroyed the computer, but it is not destroyed.

On Rappaport Hovav & Levin’s approach, result verbs contrast with verbs that encode a manner of carrying out an action, but do not specify that any result state follows from that action. For instance, while wiping events often result in some surface becoming cleaner, the becoming cleaner is not part of the lexical entailments of the verb wipe, as it is actually a pragmatic inference drawn from speakers’ knowledge of the world when it comes to wiping events (Rappaport Hovav & Levin 2010). Evidence that supports this comes from observing that sentences like Kim just wiped the table, but the table is just as dirty as it was before or I scrubbed the tub for hours, but it didn’t get any cleaner (Rappaport Hovav & Levin 2010: 22) are not contradictory. In other words, as Rappaport Hovav & Levin (2010: 22) discuss in detail, verbs like wipe “lexically specify […] actions [which] are typically used with the intention of removing stuff from a surface, and in particular contexts, this removal will be strongly implicated; however, since it can be explicitly denied, it is not lexically encoded […] in the verb.” Manner verbs therefore should not generate contradictions when subject to something is different about x diagnostic as proposed by Beavers (2011). This is because these verbs do not encode any result state from the action they lexically specify and therefore it should be possible to explicitly deny that any result state has occurred, despite the fact that it may be (strongly) implicated in some contexts, as Rappaport Hovav & Levin discuss. This is illustrated below for the manner verbs wipe, hit and work out.

    1. (7)
    1. a.
    1. Kim just wiped the table, but nothing is different about it.
    1.  
    1. b.
    1. Kim just hit the wall, but nothing is different about it.
    1.  
    1. c.
    1. Kim just worked out hard, but nothing is different about her.

As mentioned before, Spanish and Catalan reflexive verbs that appear in the intransitive construction at stake all encode a result state that comes about accidentally as a result of an action the referent of the subject is engaged in. Some verbs in this respect involve matarse ‘kill’, electrocutarse ‘electrocute’, herirse ‘hurt’, lesionarse ‘injure’, among others, which we call matarse verbs in a broad sense and for descriptive ease. These verbs qualify as result verbs on Rappaport Hovav & Levin’s view since explicitly denying that nothing is different about the patient generates clear contradictions.

    1. (8)
    1. a.
    1. María
    2. María
    1. se
    2. se
    1. ha
    2. have.3sg.prs
    1. electrocutado,
    2. electrocute.ptcp,
    1. #pero
    2.   but
    1. no
    2. no
    1. hay
    2. there.is.3sg.prs
    1. nada
    2. nothing
    1. diferente
    2. different
    1. en
    2. in
    1. ella.
    2. her
    1. Intended: “María just got herself electrocuted, but there’s nothing different about her.”
    1.  
    1. b.
    1. Juan
    2. Juan
    1. se
    2. se
    1. ha
    2. have.3sg.prs
    1. lesionado
    2. injure.ptcp
    1. en
    2. in
    1. el
    2. the
    1. partido,
    2. game,
    1. #pero
    2.   but
    1. no
    2. no
    1. hay
    2. there.is.3sg.prs
    1. nada
    2. nothing
    1. diferente
    2. different
    1. en
    2. in
    1. él.
    2. him
    1. Intended: “Juan just got himself injured in the game, but there’s nothing different about him.”
    1.  
    1. c.
    1. Juan
    2. Juan
    1. se
    2. se
    1. ha
    2. have.3sg.prs
    1. matado
    2. kill.ptcp
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche,
    2. car,
    1. #pero
    2.   but
    1. no
    2. no
    1. hay
    2. there.is.3sg.prs
    1. nada
    2. nothing
    1. diferente
    2. different
    1. en
    2. in
    1. él.
    2. him
    1. Intended: “Juan just got himself killed in a car accident, but there’s nothing different about him.”

Crucial evidence for the generalization that we wish to advance here comes from considering other types of reflexive verbs like lavarse ‘wash’, which at first blush appear to be quite similar to matarse verbs. In this respect, an anonymous reviewer correctly points out that reflexive verbs of the lavarse sort do not appear in the intransitive construction at stake, since the following example involving lavarse can only have a run-of-the-mill reflexive interpretation where the referent of the subject carries out a washing action on themselves.

    1. (9)
    1. El
    2. the
    1. niño
    2. kid
    1. se
    2. refl
    1. ha
    2. have.3sg.prs
    1. lavado
    2. wash.ptcp
    1. en
    2. in
    1. la
    2. the
    1. ducha.
    2. shower
    1. Transitive reflexive: “The kid has washed himself in the shower.”
    2. Intransitive use: “#The kid got himself washed in the shower.”

On the present account, the fact that a verb like lavarse only has a reflexive interpretation is actually accounted for, since such a verb appears to behave like a canonical manner verb. In other words, above we proposed that for a verb to participate in the intransitive construction, it needs to encode a result state that has to come about accidentally as a result of the action the referent of the subject is engaged in. Yet, a reflexive verb of the lavarse sort does not appear to encode any result state when subject to relevant diagnostics as those developed by Rappaport Hovav & Levin (2010); Beavers (2011). Intuitively, while it is true that washing events often result in something or someone becoming cleaner, this does not appear to be a lexical entailment of a Romance verb like lavarse, since such a result state can be explicitly denied in both Spanish (10) and Catalan (11) (cf. Rappaport Hovav & Levin 2010: 22).

    1. (10)
    1. a.
    1. El
    2. the
    1. chico
    2. boy
    1. se
    2. refl
    1. ha
    2. have.3sg.prs
    1. lavado
    2. wash.ptcp
    1. en
    2. in
    1. la
    2. the
    1. ducha,
    2. shower,
    1. pero
    2. but
    1. continua
    2. continue.3sg.prs
    1. igual
    2. same
    1. de
    2. of
    1. sucio.
    2. dirty
    1. “The boy has washed himself in the shower, but he is just as dirty as before.”
    1.  
    1. b.
    1. El
    2. the
    1. chico
    2. boy
    1. ha
    2. have.3sg.prs
    1. lavado
    2. wash.ptcp
    1. la
    2. the
    1. ropa,
    2. clothes,
    1. pero
    2. but
    1. continua
    2. continue.3sg.prs
    1. igual
    2. same
    1. de
    2. of
    1. sucia.
    2. dirty
    1. “The boy has washed the clothes, but they are just as dirty as before.”
    1. (11)
    1. a.
    1. El
    2. the
    1. noi
    2. boy
    1. s’
    2. refl
    1. ha
    2. have.3sg.prs
    1. rentat
    2. wash.ptcp
    1. a
    2. in
    1. la
    2. the
    1. dutxa,
    2. shower,
    1. però
    2. but
    1. continua
    2. continue.3sg.prs
    1. igual
    2. same
    1. de
    2. of
    1. brut.
    2. dirty
    1. “The boy has washed himself in the shower, but he is just as dirty as before.”
    1.  
    1. b.
    1. El
    2. the
    1. noi
    2. boy
    1. ha
    2. have.3sg.prs
    1. rentat
    2. wash.ptcp
    1. la
    2. the
    1. roba,
    2. clothes,
    1. però
    2. but
    1. continua
    2. continue.3sg.prs
    1. igual
    2. same
    1. de
    2. of
    1. bruta.
    2. dirty
    1. “The boy has washed the clothes, but they are just as dirty as before.”

We therefore predict that reflexive verbs in Spanish and Catalan can appear in the intransitive construction iff they encode a result state that comes about accidentally as a result of an action the referent of the subject is engaged in. This is further illustrated below, where a reflexive verb like herirse ‘hurt’, which qualifies as a result verb when subject to relevant result diagnostics (12-a), does appear in the intransitive construction (12-b), as well as showing a run-of-the-mill reflexive interpretation in the right contexts (12-c).

    1. (12)
    1. a.
    1. María
    2. María
    1. se
    2. se
    1. ha
    2. have.3sg.prs
    1. herido,
    2. hurt.ptcp,
    1. #pero
    2.   but
    1. no
    2. no
    1. hay
    2. there.is.3sg.prs
    1. nada
    2. nothing
    1. diferente
    2. different
    1. en
    2. in
    1. ella.
    2. her
    1. Intended: “María just got herself hurt, but there’s nothing different about her.”
    1.  
    1. b.
    1. Los
    2. the
    1. bomberos
    2. firefighters
    1. rescatan
    2. rescue.3pl.prs
    1. a
    2. dom
    1. una
    2. a
    1. mujer
    2. women
    1. que
    2. that
    1. se
    2. se
    1. hirió
    2. hurt.3sg.pfv
    1. cuando
    2. when
    1. realizaba
    2. do.3sg.ipfv
    1. una
    2. a
    1. ruta.
    2. route
    1. “The firefighters rescue a woman that got herself hurt when she was hiking.” (Web)
    1.  
    1. c.
    1. Mattie
    2. Mattie
    1. Larson
    2. Larson
    1. se
    2. refl
    1. hirió
    2. hurt.3sg.pfv
    1. para
    2. to
    1. evitar
    2. avoid.ifv
    1. entrenar
    2. train.ifv
    1. con
    2. with
    1. Nassar.
    2. Nassar
    1. “Mattie Larson hurt himself in order to avoid having to train with Nassar.” (Web)

A final piece of evidence for the present generalization regarding reflexive verbs in Spanish and Catalan comes from considering a reflexive verb like asesinarse ‘murder’, which does entail a result state of being dead but crucially requires the referent of the subject to act intentionally (Talmy 1975; Dowty 1991; Van Valin & Wilkins 1996; Rooryck & Wyngaerd 2011; Grano 2016; Solstad & Bott 2017; Tubino-Blanco 2020). A verb like asesinarse therefore contrasts with matarse ‘kill’, which entails the same result state yet it does not impose any intentionality requirement. This is illustrated below.

    1. (13)
    1. a.
    1.   El
    2.   the
    1. medico
    2. doctor
    1. mató
    2. kill.3sg.pfv
    1. al
    2. dom=the
    1. paciente
    2. patient
    1. accidentalmente.
    2. accidentally
    1.   “The doctor killed the patient by accident.”
    1.  
    1. b.
    1. #El
    2.   the
    1. terrorista
    2. terrorist
    1. asesinó
    2. murder.3sg.pfv
    1. al
    2. dom=the
    1. presidente
    2. president
    1. accidentalmente.
    2. accidentally
    1.   Intended: “The terrorist murdered the president by accident.”

The fact that a verb like asesinarse requires the referent of the subject to act intentionally when bringing about the result state it encodes explains why this reflexive verb cannot appear in the intransitive construction, as asesinarse and the intransitive construction have contradictory entailments.2

    1. (14)
    1. a.
    1.   Haider
    2.   Haider
    1. conducía
    2. drive.3sg.ipfv
    1. ebrio
    2. drunk
    1. cuando
    2. when
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche.
    2. car.
    1.   “Haider got himself killed as he was driving under the influence.” (Web)
    1.  
    1. b.
    1. #Haider
    2.   Haider
    1. conducía
    2. drive.3sg.ipfv
    1. ebrio
    2. drunk
    1. cuando
    2. when
    1. se
    2. se
    1. asesinó
    2. murder.3sg.pfv
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche.
    2. car.
    1.   Intended: “Haider got himself murdered as he was driving under the influence.” (Web)

3.1 Halfway between agents and patients

In this section, we observe that the subjects of matarse verbs in Spanish and Catalan, when found in the intransitive construction, appear to have both internal and external argument properties. In order to make our case, we observe intransitive uses of matarse verbs have distinct syntactic and semantic properties than reflexive uses, despite sharing the same surface form.

3.1.1 Semantic properties

Similarly to what Pineda & Berro (2020: 1) note for a class of agentive verbs in Basque, the subject of matarse verbs in the intransitive construction is “interpreted as both the force initiating the process and the entity undergoing it.” The claim that the subject’s referent needs to be the force initiating the process, i.e., it has external argument properties, for the construction to be felicitous is illustrated below in (15). In this case, the context makes it clear that the only possible thematic interpretation of the subject is that of a canonical patient, i.e., an argument that simply undergoes change. Consequently, matarse verbs cannot felicitously describe this scenario since the subject’s referent is not engaged in an action that ultimately ends up killing them. Only verbs whose subjects involve canonical patient arguments such as morir ‘die’ can (see Tubino-Blanco 2020 for similar contrasts).

    1. (15)
    1. Context: Juan was unconscious on the ground, a big stone fell on his head, so …
    2. Juan #se mató/OKmurió en el suelo.
    3. ‘Juan #got himself killed/died on the ground.’

Crucially, though, if the contexts allow for a reading in which the subject’s referent is indeed interpreted as the force initiating the process, then matarse verbs can felicitously describe them, even if they are quite similar to those in (15). Compare this in the example below.

    1. (16)
    1. Context: Juan was climbing without a security rope when he fell, hit his head on the ground, and lay unconscious for a while before he died.
    2. Juan se mató escalando.
    3. ‘Juan got himself killed when climbing.’

The contrast illustrated in (15) and (16) therefore provides evidence that the subject’s referent of matarse verbs in the intransitive construction needs to be engaged in an action that ultimately leads to the change it undergoes, e.g., in (16) the subject’s referent is the force initiating a climbing activity that ultimately leads to its accidental death. In other words, the subject’s referent needs to hold some degree of responsibility for what occurs to them.3

Evidence for the claim that the subject also has internal argument properties, i.e., it qualifies as a patient in some respects, comes from various sources. First, it appears clear that the subject behaves as a canonical patient in that the subject’s referent is the entity that undergoes the change of state that the verbs encode, as discussed above (8). This is further illustrated in the example below, in which explicitly denying this generates a clear contradiction (Rappaport Hovav & Levin 2010; Beavers & Koontz-Garboden 2012).

    1. (17)
    1. Juan
    2. Juan
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche
    2. car
    1. …
    2. …
    1. #pero
    2. but
    1. no
    2. no
    1. murió.
    2. die.3sg.pfv
    1. Intended: “Juan got himself killed in a car accident … but he didn’t die.”

More importantly, we observe that intransitive uses of matarse verbs show a striking syntactic and semantic contrast with transitive reflexives, despite sharing the same surface form in Spanish and Catalan.4 Only reflexives involve canonical agents, i.e., an argument that acts intentionally, whereas intransitive uses entail accidental death, electrocution etc. We illustrate this below with Catalan examples which all explicitly specify that death, electrocution or hurting is an accidental result from the action the subjects’ referents are engaged in. For instance, it is quite clear that in (18-a) electrocution is an unforeseen consequence of the stealing activity. Similar comments apply to the rest of the examples.

    1. (18)
    1. a.
    1. En
    2. in
    1. estat
    2. state
    1. greu
    2. grave
    1. un
    2. a
    1. home
    2. man
    1. que
    2. that
    1. es
    2. se
    1. va
    2. go.3sg.prs
    1. electrocutar
    2. electrocute.ifv
    1. quan
    2. when
    1. intentava
    2. try.3sg.ipfv
    1. robar
    2. steal.ifv
    1. coure
    2. copper
    1. a
    2. in
    1. Cambrils.
    2. Cambrils
    1. “A man is severely injured after he got himself electrocuted when he was trying to steal copper in Cambrils.” (Web)
    1.  
    1. b.
    1. La
    2. the
    1. qüestió
    2. question
    1. és
    2. is
    1. que
    2. that
    1. l’
    2. the
    1. home
    2. man
    1. no
    2. no
    1. es
    2. se
    1. va
    2. go.3sg.prs
    1. matar
    2. kill.ifv
    1. quan
    2. when
    1. va
    2. go.3sg.prs
    1. caure.
    2. fall.ifv
    1. “The point is that the man didn’t get himself killed when he fell.” (Web)
    1.  
    1. c.
    1. Rescatada
    2. rescue.ptcp.f.sg
    1. una
    2. a
    1. dona
    2. woman
    1. que
    2. that
    1. es
    2. se
    1. va
    2. go.3sg.prs
    1. ferir
    2. hurt.ifv
    1. quan
    2. when
    1. realitzava
    2. do.3sg.ifv
    1. senderisme
    2. hiking
    1. a
    2. in
    1. Calp.
    2. Calp
    1. “A woman who got herself hurt was rescued when she was hiking in Calp.” (Web)

This lack of intentionality is explicitly illustrated below (Spanish examples from the Web), where intransitive uses are naturally compatible with adverbial expressions such as sin querer ‘by accident’ or accidentalmente ‘accidentally’. This is relevant for the present purposes since these expressions are incompatible with interpretations where the subjects’ referents intentionally cause the change that they undergo.

    1. (19)
    1. a.
    1. La
    2. the
    1. misma
    2. same
    1. chica
    2. girl
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. sin
    2. without
    1. querer,
    2. want.inf
    1. al
    2. at.the
    1. mezclar
    2. mix.ifv
    1. bebidas
    2. drinks
    1. alcoholicas.
    2. alcoholic
    1. “The same girl got herself killed by accident as she mixed alcoholic drinks.”
    1.  
    1. b.
    1. Un
    2. a
    1. hombre
    2. man
    1. se
    2. se
    1. hirió
    2. hurt.3sg.pfv
    1. accidentalmente
    2. accidentally
    1. con
    2. with
    1. el
    2. the
    1. arma
    2. gun
    1. que
    2. that
    1. usaba
    2. use.3sg.ipfv
    1. para
    2. to
    1. cazar.
    2. hunt
    1. “A man got himself hurt accidentally with the gun he used to hunt.”
    1.  
    1. c.
    1. El
    2. the
    1. hombre
    2. man
    1. se
    2. se
    1. electrocutó
    2. electrocute.3sg.pfv
    1. accidentalmente
    2. accidentally
    1. al
    2. at.the
    1. afeitarse
    2. shave.ifv=refl
    1. en
    2. in
    1. la
    2. the
    1. ducha.
    2. shower
    1. “The man got himself electrocuted by accident when he was shaving in the shower.”

In contrast, the same adverbial expressions generate clear contradictions with reflexive uses. This is expected since we suggest that the subject of reflexives involves a canonical agent argument that needs to act intentionally, as we discuss in detail in Section 4. Insofar as the interpretation generated by reflexives is that of e.g., committing suicide in the case of matarse, the fact that the subjects in this case need to qualify as intentional agents is therefore expected, since suicide is intentional by definition.

    1. (20)
    1. a.
    1. #El
    2.   the
    1. sospechoso
    2. suspect
    1. del
    2. of.the
    1. homicidio
    2. homicide
    1. se
    2. refl
    1. mató
    2. kill.3sg.pfv
    1. accidentalmente
    2. accidentally
    1. al
    2. at.the
    1. estar
    2. be.ifv
    1. rodeado
    2. surrounded
    1. por
    2. by
    1. la
    2. the
    1. policía.
    2. police
    1.   Intended: “The suspect killed himself accidentally when he was surrounded by the police.”
    1.  
    1. b.
    1. #Mattie
    2.   Mattie
    1. Larson
    2. Larson
    1. se
    2. refl
    1. hirió
    2. hurt.3sg.pfv
    1. accidentalmente
    2. accidentally
    1. para
    2. to
    1. evitar
    2. avoid.ifv
    1. entrenar
    2. train.ifv
    1. con
    2. with
    1. Nassar.
    2. Nassar
    1.   Intended: “Mattie Larson hurt himself accidentally in order to avoid having to train with Nassar.”
    1.  
    1. c.
    1. #El
    2.   the
    1. hombre
    2. man
    1. que
    2. that
    1. se
    2. refl
    1. electrocutó
    2. electrocute.3sg.pfv
    1. accidentalmente
    2. accidentally
    1. para
    2. to
    1. probar
    2. test
    1. una
    2. a
    1. terapia
    2. therapy
    1. genética
    2. genetic
    1. antienvejecimiento.
    2. anti-aging
    1.   Intended: “The man that electrocuted himself accidentally to test an antigaging genetic therapy.”

3.1.2 Syntactic properties

Further evidence for the claim that only the subject of matarse verbs in intransitive constructions has internal argument properties comes from considering the possibility of bare plural subjects appearing in postverbal position. This is relevant for the present purposes since this is often taken to be a canonical unaccusativity diagnostic that probes for the presence of internal arguments (e.g., Perlmutter 1978 for English, Burzio 1986 for Italian, Rigau 1990; Mateu & Massanell i Messalles 2015 for Catalan or Cifuentes Honrubia 1999 for Spanish). Namely, in Romance languages, it has been observed that only bare plural subjects of unaccusative verbs, i.e., those verbs whose subject is an internal argument, can naturally appear as unmarked in postverbal position (21-a)-(22-a), in contrast to bare plural subjects of unergative verbs, which by virtue of involving an external argument, resist appearing postverbally (21-b)-(22-b), or involve highly marked constructions at best. This is illustrated below for Spanish and Catalan respectively.

    1. (21)
    1. a.
    1.     Llegaron
    2.     arrive.3pl.pfv
    1. trenes.
    2. trains
    1.     “There arrived trains.”
    1.  
    1. b.
    1. ??Trabajaron
    2.     work.3pl.pfv
    1. personas.
    2. people
    1.     Intended: “There worked people.”
    1. (22)
    1. a.
    1.     Neixen
    2.     be.born.3pl.prs
    1. nens
    2. children
    1. cada
    2. every
    1. dia.
    2. day
    1.     “Children are born every day.”
    1.  
    1. b.
    1. ??Riuen
    2.     laugh.3pl.prs
    1. nens
    2. children
    1. cada
    2. every
    1. dia.
    2. day.
    1.     Intended: “There laugh children every day.”

In this respect, we observe that only bare plural subjects of intransitive uses of matarse verbs can naturally appear postverbally (23), in contrast to bare plural subjects of reflexive uses (24), which resist appearing in this position.5

    1. (23)
    1. a.
    1. Todos
    2. all
    1. los
    2. the
    1. días
    2. days
    1. se
    2. se
    1. matan
    2. kill.3pl.prs
    1. motociclistas
    2. motorcyclists
    1. […]
    2. […]
    1. por
    2. for
    1. la
    2. the
    1. imprudencia
    2. recklessness
    1. de
    2. of
    1. estos
    2. these
    1. señores.
    2. men
    1. “Everyday there are motorcyclists that get themselves killed because of their recklessness.” (CE)
    1.  
    1. b.
    1. El
    2. the
    1. equipo
    2. team
    1. se
    2. se
    1. encuentra
    2. find.3sg.prs
    1. muy
    2. very
    1. bien,
    2. well,
    1. pero
    2. but
    1. infortunadamente
    2. unfortunately
    1. se
    2. se
    1. lesionaron
    2. injure.3pl.pfv
    1. jugadoras
    2. players
    1. como
    2. like
    1. Cindy
    2. Cindy
    1. Ramírez.
    2. Ramírez
    1. “The team is really well, but unfortunately some players like Cindy Ramírez got themselves injured.” (CE)
    1.  
    1. c.
    1. Se
    2. se
    1. electrocutan
    2. electrocute.3pl.prs
    1. trabajadores
    2. workers
    1. que
    2. that
    1. hacían
    2. do.3pl.ipfv
    1. reparaciones
    2. repairs
    1. en
    2. at
    1. un
    2. a
    1. comercio.
    2. shop
    1. “Some workers got themselves electrocuted when they were making some repairs at a shop.” (Web)
    1. (24)
    1. a.
    1. ??Para
    2.     to
    1. evitar
    2. avoid.ifv
    1. ser
    2. be.ifv
    1. capturados
    2. capture.ptcp.m.pl
    1. por
    2. by
    1. el
    2. the
    1. enemigo,
    2. enemy,
    1. se
    2. refl
    1. mataron
    2. kill.3pl.pfv
    1. soldados
    2. soldiers
    1. americanos.
    2. american
    1.     “In order to avoid being captured by the enemy, American soldiers killed themselves.”
    1.  
    1. b.
    1. ??Para
    2.     to
    1. no
    2. no
    1. entrenar
    2. train.ifv
    1. con
    2. with
    1. lluvia,
    2. rain,
    1. se
    2. refl
    1. lesionaron
    2. hurt.3pl.pfv
    1. jugadores
    2. players
    1. como
    2. like
    1. Messi
    2. Messi
    1. o
    2. or
    1. Neymar.
    2. Neymar
    1.     “In order to avoid training while it rained, players like Messi or Neymar injured themselves.”
    1.  
    1. c.
    1. ??Con
    2.     with
    1. el
    2. the
    1. fin
    2. goal
    1. de
    2. of
    1. cobrar
    2. collect
    1. el
    2. the
    1. seguro,
    2. insurance,
    1. se
    2. refl
    1. electrocutaron
    2. electrocute.3pl.pfv
    1. trabajadores.
    2. workers
    1.     “In an attempt to collect the insurance money, some workers electrocuted themselves.”

Last, an additional piece of evidence for the claim that the subjects of reflexives and intransitive uses have distinct argument properties comes from considering en cliticization, which involves a patient interpretation of the subject in Catalan. As is well-known from the literature (Rigau 1990; 1991; 1994; Rosselló 2008; Acedo-Matellán & Mateu 2015), en is a Catalan pronoun that picks out indefinite internal arguments of transitive (25) and unaccusative verbs (26), but never external arguments of unergative verbs (27) (see Burzio 1986; Levin & Rappaport Hovav 1995 for Italian).

    1. (25)
    1. a.
    1. El
    2. the
    1. Joan
    2. Joan
    1. ha
    2. have.3sg.prs
    1. comprat
    2. buy.ptcp
    1. patates.
    2. potatoes
    1. “Joan has bought potatoes.”
    1.  
    1. b.
    1. El
    2. the
    1. Joan
    2. Joan
    1. n’
    2. en
    1. ha
    2. have.3sg.prs
    1. comprat,
    2. buy.ptcp,
    1. de
    2. of
    1. patates.
    2. potatoes
    1. “Joan has indeed bought potatoes.”
    1. (26)
    1. a.
    1. Han
    2. have.3sg.prs
    1. arribat
    2. arrive.ptcp
    1. estudiants
    2. students
    1. italians
    2. Italian
    1. aquest
    2. this
    1. any.
    2. year
    1. “Italian students have arrived this year.”
    1.  
    1. b.
    1. N’
    2. en
    1. han
    2. have.3sg.prs
    1. arribat,
    2. arrive.ptcp,
    1. d’
    2. of
    1. estudiants
    2. students
    1. italians
    2. Italian
    1. aquest
    2. this
    1. any.
    2. year
    1. “When it comes to Italian students, there have arrived some this year.”
    1. (27)
    1. a.
    1.   Molts
    2.   lots
    1. nens
    2. kids
    1. han
    2. have.3pl.prs
    1. jugat
    2. play.ptcp
    1. avui.
    2. today
    1.   “A lot of children have played today.”
    1.  
    1. b.
    1. *N’
    2.   en
    1. han
    2. have.3pl.prs
    1. jugat
    2. play.ptcp
    1. molts,
    2. lots,
    1. de
    2. of
    1. nens
    2. children
    1. avui.
    2. today
    1.   Intended: “When it comes to children, there have played some today.”

Insofar as en cliticization in Catalan always picks out internal arguments, and never external ones, as evidenced by the fact that the subjects of unergative verbs cannot be pronominalized by en (27), an additional prediction arises. Namely, only the subjects of matarse verbs in the intransitive construction should be compatible with cliticization of this sort. As illustrated by the Catalan examples below, this prediction appears to be borne out.

    1. (28)
    1. a.
    1. Ja
    2. already
    1. no
    2. no
    1. podem
    2. be.able.1pl.prs
    1. recórrer
    2. resort.ifv
    1. al
    2. to.the
    1. tòpic,
    2. cliche,
    1. fins
    2. until
    1. que
    2. that
    1. algú
    2. somebody
    1. es
    2. se
    1. mati,
    2. kill.3sg.prs.sbjv,
    1. perquè
    2. because
    1. malauradament
    2. sadly
    1. ja
    2. already
    1. se
    2. se
    1. n’
    2. en
    1. han
    2. have.3pl.prs
    1. matat
    2. kill.ptcp
    1. molts.
    2. lots
    1. “Sadly, lots of people already got themselves killed, so we can no longer say this won’t stop until somebody gets themselves killed.” (Web)
    1.  
    1. b.
    1. Els
    2. the
    1. osonencs
    2. people.from.Osona
    1. […]
    2. […]
    1. acostumats
    2. accustom.ptcp.m.pl
    1. a
    2. to
    1. anar
    2. go.inf
    1. amb
    2. with
    1. el
    2. the
    1. Simca1200
    2. Simca1200
    1. a
    2. to
    1. fons
    2. bottom
    1. quan
    2. when
    1. va
    2. go.3sg.prs
    1. sortir
    2. leave.ifv
    1. el
    2. the
    1. R5
    2. R5
    1. turbo
    2. turbo
    1. se’
    2. se
    1. n
    2. en
    1. van
    2. go.3pl.prs
    1. matar
    2. kill.ifv
    1. molts.
    2. lots
    1. “Some people from Osona got themselves killed when the R5 turbo got released as they were used to the Simca1200.” (Web)

In contrast, reflexive uses of the same verbs should resist it, as en cliticization involves a patient interpretation of their subjects. As illustrated by the examples below, this also appears to be borne out.6

    1. (29)
    1. a.
    1.    Durant
    2.    during
    1. la
    2. the
    1. primera
    2. first
    1. guerra
    2. war
    1. mundial,
    2. world,
    1. molts
    2. lots
    1. soldats
    2. soldiers
    1. es
    2. refl
    1. mataven
    2. kill.3pl.ipfv
    1. per
    2. to
    1. evitar
    2. avoid.inf
    1. ser
    2. be.ifv
    1. capturats
    2. capture.ptcp.m.pl
    1. per
    2. by
    1. l’
    2. the
    1. enemic.
    2. enemy
    1.    “During the first world war, a lot of soldiers killed themselves in order to avoid being captured by the enemy.”
    1.  
    1. b.
    1. ??Durant
    2.    during
    1. la
    2. the
    1. primera
    2. first
    1. guerra
    2. war
    1. mundial,
    2. world,
    1. se
    2. refl
    1. ‘n
    2. en
    1. mataven
    2. kill.3pl.ipfv
    1. molts,
    2. lots,
    1. de
    2. of
    1. soldats,
    2. soldiers,
    1. per
    2. to
    1. evitar
    2. avoid.ifv
    1. ser
    2. be.ifv
    1. capturats
    2. capture.ptcp.m.pl
    1. per
    2. by
    1. l’
    2. the
    1. enemic.
    2. enemy
    1. (30)
    1. a.
    1.    Quan
    2.    when
    1. diagnostiquen
    2. diagnose.3pl.sg
    1. un
    2. a
    1. cancer
    2. cancer
    1. terminal,
    2. terminal,
    1. moltes
    2. lots
    1. persones
    2. people
    1. es
    2. refl
    1. maten
    2. kill.3pl.prs
    1. per
    2. to
    1. evitar
    2. avoid.ifv
    1. sofrir.
    2. suffer.ifv
    1.    “When people are diagnosed with a terminal cancer, many kill themselves.”
    1.  
    1. b.
    1. ??Quan
    2.    when
    1. diagnostiquen
    2. diagnose.3pl.sg
    1. un
    2. a
    1. càncer
    2. cancer
    1. terminal,
    2. terminal,
    1. se’
    2. refl
    1. n
    2. en
    1. maten
    2. kill.3pl.prs
    1. moltes,
    2. lots,
    1. de
    2. of
    1. persones,
    2. people,
    1. per
    2. to
    1. evitar
    2. avoid.ifv
    1. sofrir.
    2. suffer.ifv

3.2 Summary

In this section, we have provided evidence for the claim that the subjects of matarse verbs in Spanish and Catalan when found in what we have called an intransitive construction have both internal and external argument properties. In order to make our claim, we have further considered reflexive uses of the same verbs, since they share the same surface form. We have provided evidence that these two constructions involve distinct syntactic and semantic properties by considering a battery of semantic and syntactic diagnostics that probe for the presence of external and internal arguments. Similarly, the very fact that examples like Spanish Juan se mató or Catalan El Joan es va matar are ambiguous between a reflexive and an intransitive interpretation strongly suggests that there are two distinct argument structures that can be associated with the same linear form, as we discuss in detail in the next section.

Semantically, the subjects of intransitive uses show external argument properties since their referents need to be engaged in an action that ultimately leads to the change they undergo. Namely, the subjects’ referents need to hold some degree of responsibility for what ultimately occurs to them as they are the force that initiates the process that they themselves undergo (Pineda & Berro 2020). We have observed that this is not a requirement for verbs that involve canonical patient arguments such as unaccusative verbs of the morir ‘die’ sort (15). Similarly, only the subjects of reflexives qualify as canonical agents that act intentionally in causing the change encoded by the verb, in contrast to intransitive uses, which require the result of the event to arise accidentally. Syntactically, we have observed that only the subjects of intransitive uses, in contrast to the subjects of reflexives, behave as canonical internal arguments as they can naturally appear postverbally and be pronominalized by the pronoun en in Catalan.

We have therefore arrived at the conclusion that only the subjects of reflexives constitute canonical external arguments, whereas the subjects of intransitive uses have mixed external and internal argument properties. Syntactically they behave as internal arguments, but semantically they require that their referent be actively engaged in the action denoted by the event, which is an interpretation characteristic of external arguments. In the next section, following Pineda & Berro (2020), as well as Schäfer (2008) and Alexiadou et al. (2015), we lay out analyses that account for such differences in the grammatical properties of these two distinct constructions rooted in the claim that subjects exhibit distinct syntactic and semantic properties depending on the position in the syntax they are generated in (Alexiadou 2014; Tollan 2018; Tollan & Oxford 2018).

4 Analysis

We adopt a syntactic decompositional view of verb meaning and event structure in which verbs are created in the syntax by combining lexical roots with functional heads (von Stechow 1996; Marantz 1997; Mateu 2002; Borer 2003; 2005; Harley 2003; 2005; 2013; Pylkkännen 2008; Ramchand 2008; Mateu & Acedo-Matellán 2012; Acedo-Matellán & Mateu 2014; Acedo-Matellán 2016, among many others). We assume the existence of a little v head with eventive semantics (Marantz 1997; 2013; Harley 2005; 2013; Pylkkännen 2008; Acedo-Matellán 2016; Wood & Marantz 2017, among others), and a Pred head (Bowers 1993) which functions as a relator responsible for structuring small clauses that serve as complement to the v head in predicates involving change of state or location (Hoekstra 1988; 1992). The small clause introduces the undergoer of the event of change and the final state or location in the specifier and complement of Pred respectively. Agent external arguments are instead introduced by the functional head Voice (Kratzer 1996). The following predicate, denoting a resultative event of change of state, namely, an event of becoming broken, undergone by a vase, illustrates this.

    1. (31)
    1. John broke a vase.

Following Pineda & Berro (2020), we propose that the hybrid behavior of the subjects of matarse verbs can be explained by appealing to the claim that subjects can be assigned more than one distinct semantic role by virtue of being associated with more than one distinct head in the syntax. Specifically, adopting Pineda & Berro’s analysis of a class of agentive verbs in Basque, we argue that the subjects of matarse verbs in the intransitive construction initially merge vP internally, in the specifier position of a small clause subpredicate, and are co-indexed with the unpronounced specifier of a thematic, but syntactically intransitive Voice head (Schäfer 2008; Alexiadou et al. 2015).7 We assume that the Voice head is phonologically realized by means of a se/es morpheme that is phonologically homonymous with the reflexive pronoun se/es (Fábregas 2021), therefore giving rise to ambiguity in the interpretation of surface forms like Spanish Juan se mató as either involving an intransitive, i.e., John got himself killed, or a reflexive use, i.e., John killed himself, of the verb.8

More specifically, subjects with a hybrid behavior have both internal and external argument properties because Voice assigns its agent argument to the referent of the internal argument. Assuming the analysis laid out in Pineda & Berro (2020), co-indexation therefore makes it possible for these subjects to bear two semantic roles, as they are interpreted as arguments of both Voice and the small clause.9

    1. (32)
    1. Haider
    2. Haider
    1. […]
    2.  
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche.
    2. car.
    1. “Haider got himself killed in a car accident.”

In contrast, reflexive uses of matarse verbs involve a transitive configuration where subjects are externally merged in the specifier of the Voice head, while a reflexive pronoun (e.g., se in Spanish,) co-referent with the subject, occupies the internal argument position (Doron & Rappaport Hovav 2009; Sportiche 2014; 2020; Martin et al. 2023).

    1. (33)
    1. El
    2. the
    1. sospechoso
    2. suspect
    1. […]
    2.  
    1. se
    2. refl
    1. mató
    2. kill.3sg.pfv
    1. al
    2. at.the
    1. estar
    2. be.ifv
    1. rodeado
    2. surrounded
    1. por
    2. by
    1. la
    2. the
    1. policía.
    2. police
    1. “The suspect killed himself when he was surrounded by the police.”

We close by providing an additional piece of evidence for the claim that reflexives and intransitive uses have distinct grammatical properties. This comes from considering overt reflexives, e.g., oneself. In this respect, we observe that only reflexive uses of matarse verbs naturally combine with explicit reflexive expressions such as Spanish a sí mismo, whereas intransitive uses do not. Compare this below.

    1. (34)
    1. a.
    1. ??Juan
    2.    Juan
    1. se
    2. se
    1. mató
    2. kill.3sg.pfv
    1. a
    2. dom
    1. sí
    2. refl
    1. mismo
    2. same
    1. en
    2. in
    1. un
    2. a
    1. accidente
    2. accident
    1. de
    2. of
    1. coche.
    2. car
    1.    “Juan got himself killed in a car accident.”
    1.  
    1. b.
    1.    Keith
    2.    Keith
    1. Young
    2. Young
    1. […]
    2.  
    1. se
    2. refl
    1. ha
    2. have.3.sg.prs
    1. matado
    2. kill.ptcp
    1. a
    2. dom
    1. sí
    2. refl
    1. mismo
    2. same
    1. y
    2. and
    1. a
    2. dom
    1. sus
    2. his
    1. cuatro
    2. four
    1. hijos
    2. children
    1. pequeños.
    2. little
    1.    “Keith young has killed himself and his four little kids.” (Web)

The contrast in (34) naturally follows if intransitive uses of matarse verbs involve the intransitive configuration in (32), in contrast to reflexives which involve the transitive one in (33). Assuming, following Uriagereka (1988; 2001); Torrego (1994); Cecchetto (2000); Belletti (2005), among others, that the doubling construction in (34-b) arises via an adjunct/specifier relation between the explicit reflexive expressions (e.g., a sí mismo) and the reflexive clitic se, the construction is licensed in the transitive reflexive use but not in the intransitive use because only in the transitive reflexive use, the clitic is merged in a phrasal position, which is additionally c-commanded by the subject argument that binds the anaphoric expressions, as illustrated in (35).

    1. (35)
    1. Keith Young […] se ha matado a sí mismo.

5 Conclusion

The present paper has provided independent novel evidence from Spanish and Catalan that supports the view originally defended by Pineda & Berro (2020) regarding the existence of what they refer to as hybrid verbs. We have concerned ourselves with an intransitive use of what we have called matarse verbs in Spanish and Catalan, and have observed that the subjects of these verbs when found in an intransitive configuration show mixed-behavior interpretations between agents and patients. We have proposed that the hybrid nature of the subjects of matarse verbs can be explained by appealing to Pineda & Berro’s claim that subjects can be assigned two distinct semantic roles by virtue of being associated with more than one distinct head in the syntax. Ultimately, the data from matarse verbs in Spanish and Catalan has allowed us to show that agents and external arguments do not always correlate since some verb classes can have an identical surface form, despite involving underlyingly distinct argument alignments.

Abbreviations

The glossing in this paper follows the conventions stated in the Leipzig Glossing Rules.

Funding information

Ausensi was supported by the project PID2022-136610NB-I00 (MCIN/AEI/10.13039/501100011033/FEDER, UE). Bigolin was supported by the grant FPI PRE2018-085396 (AEI/European Social Fund), and projects PID2021-123617NB-C41 (Ministerio de Ciencia e Innovación) and 2021SGR00787 (AGAUR-Generalitat de Catalunya).

Acknowledgements

The present paper would not exist had it not been for Toni Bassaganyas-Bars, who drew the first author’s attention to the relevant examples that we have considered. We are therefore indebted to him. We are grateful to three anonymous reviewers for their comments, suggestions and criticisms, which allowed us to improve the paper as they made us think deeper about the data and the analyses. We also wish to extend our gratitude to Víctor Acedo-Matellán and Isabel Oltra-Massuet, editors of the Special Issue this paper has been published in, for inviting us to contribute to it. We also thank Johan Rooryck, our handling editor, for his help and guidance throughout the reviewing process. The present paper has benefited from discussion with Víctor Acedo-Matellán, Isabel Crespí, Josep M. Fontana, Jordi Ginebra, Jaume Mateu, Isabel Oltra-Massuet, Anna Pineda, Cristina Real-Puigdollers and Carles Royo. All remaining errors are solely ours.

Competing interests

The authors have no competing interests to declare.

Authors’ contributions

Authors are ordered alphabetically according to their surnames, therefore they should both be considered co-authors.

Notes

  1. The examples in the present paper mainly come from basic web searches (Web), from Corpus de referencia del español actual (CREA) (Real Academia Española on line) and Corpus del español NOW corpus (CE) (Davies on line). Examples with no source have been constructed by us. [^]
  2. Although the transitive reflexive use of asesinarse is not frequently attested, insofar as speakers’ knowledge of real-world events of murdering do not generally involve the referent of the subject to carry the murdering on themselves, but on an independent second participant, there are still naturally-occurring examples of this sort, both in English and Spanish: e.g., There is one argument for the view that Socrates murdered himself (Web) or Si alguna vez logramos quedar en los libros de historia, será por la destrucción masiva de una especie que se asesinó a sí misma “If we ever make the history books, it will be because of a species that murdered themselves.” (Web) [^]
  3. See Homer & Sportiche (2011) who note the same contrasts for similar constructions in French. [^]
  4. For the sake of the discussion, it is worth mentioning again that, as briefly discussed above, in Spanish and Catalan, reflexives and what we have called intransitive uses of matarse verbs have the same surface form, yet this does not appear to be the case for all languages. For instance, in English, intransitive uses and reflexive interpretations are expressed by two distinct structures: the former are most naturally translated as get oneself killed/electrocuted, whereas the latter receive a canonical reflexive paraphrase of the kill/electrocute oneself sort. This suggests that intransitive and reflexives uses in Spanish and Catalan appear to involve distinct syntactic and semantic configurations, despite sharing the same surface form, since some languages have two structures with different surface form for each interpretation. [^]
  5. The judgments of the following examples marked as ?? were collected by surveying five different speakers of Spanish and Catalan who agreed with the contrasts in acceptability. [^]
  6. For some speakers, en cliticization is possible with reflexives if the predicate involves the locative pronoun hi ‘there’. This is expected since this pronoun is known to force an existential interpretation with an unaccusative flavor. As Pineda (2020: 150) notes, this is a well-known fact as it has been discussed extensively at least since the 1970s first by Solà (1973; 1987; 1994; 1999) and later by Rigau (1992; 1997a; b). Namely, canonical unergative verbs like treballar ‘work’ in Catalan can be compatible with en in such contexts, especially if they also involve locative inversion, since it favors the unaccusative flavor even further, e.g., En aquesta fàbrica n’hi treballen moltes, de dones ‘Lots of women work in this factory’. As Pineda (2020: 150) discusses, in these examples, the unergative verb is unaccusativized as it is integrated into an unaccusative construction. [^]
  7. Even though Pineda & Berro (2020) do not provide a definition of co-indexation, following the logic of their account, one plausible definition would be to understand co-indexation as a mechanism that applies to cases in which the uninterpretable features of two distinct functional heads (in the sense of Chomsky 1995) are checked by the same argument in the absence of further operations (e.g., Move). [^]
  8. For discussion of the different uses of se in Spanish, see Armstrong (2011) and Fábregas (2021) and references therein. [^]
  9. An anonymous reviewer asks how the claim that verbs are inherently specified as encoding result states or manners of actions (Rappaport Hovav & Levin 2010) factors into the derivation of the structures into which verbs can be integrated. We do not commit ourselves to claiming that an ontological-type classification of manner or result has direct correlations in the syntax (see Alexiadou et al. 2015 precisely for this view). However, assuming Marantz’s (2005) proposal (also Acedo-Matellán 2016) that the objects of atelic, non-resultative transitive predicates based on ‘manner’ roots of the lavarse sort are merged syntactically within an adjunct of the vP, the fact that examples like Juan se lavó ‘Juan washed himself’ are not ambiguous between an intransitive and a reflexive use interpretation (cf. (9)), in contrast to examples that involve result verbs which are indeed ambiguous, e.g., Juan se mató ‘Juan killed himself/Juan got himself killed’, can be said to follow from the present approach since co-indexation in these cases does not appear to be possible. This is so because of the adjunct status of the object of manner verbs, as claimed by Marantz (2005), and adjuncts are considered an opaque domain for the probing of elements inside them (Ross 1967; Huang 1982; Boeckx 2003; Chomsky 2004: among many others). In any case, what is crucial for the present purposes is that a verb encoding a result state can appear in the intransitive construction iff the result state it encodes can come about accidentally as a direct consequence of the action the referent of the subject is engaged in, as discussed above. [^]

References

Acedo-Matellán, Víctor. 2016. The morphosyntax of transitions: A case study in Latin and other languages. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780198733287.001.0001

Acedo-Matellán, Víctor & Mateu, Jaume. 2014. From syntax to roots: A syntactic approach to root interpretation. In Alexiadou, Artemis & Borer, Hagit & Schäfer, Florian (eds.), The syntax of roots and the roots of syntax, 259–281. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199665266.003.0002

Acedo-Matellán, Víctor & Mateu, Jaume. 2015. Los verbos psicológicos: Raíces especiales en estructuras corrientes. In Marín, Rafael (ed.), Los predicados psicológicos, 81–109. Madrid: Visor Libros.

Alexiadou, Artemis. 2014. The problem with internally caused change-of-state verbs. Linguistics 52(4). 879–909. DOI:  http://doi.org/10.1515/ling-2014-0011

Alexiadou, Artemis & Anagnostopolou, Elena & Schäfer, Florian. 2015. External arguments in transitivity alternations: A layering approach. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199571949.001.0001

Armstrong, Grant. 2011. Two classes of transitive verbs: Evidence from Spanish: Georgetown University dissertation.

Beavers, John. 2011. On affectedness. Natural Language and Linguistic Theory 29(2). 335–370. DOI:  http://doi.org/10.1007/s11049-011-9124-6

Beavers, John & Koontz-Garboden, Andrew. 2012. Manner and result in the roots of verbal meaning. Linguistic Inquiry 43(3). 331–369. DOI:  http://doi.org/10.1162/LING_a_00093

Beavers, John & Koontz-Garboden, Andrew. 2020. The roots of verbal meaning. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198855781.001.0001

Belletti, Adriana. 2005. Extended doubling and the VP periphery. Probus 17. 1–35. DOI:  http://doi.org/10.1515/prbs.2005.17.1.1

Boeckx, Cedric. 2003. Islands and chains: Resumption as stranding. Amsterdam: John Benjamins. DOI:  http://doi.org/10.1075/la.63

Borer, Hagit. 2003. Exo-skeletal vs. endo-skeletal explanations: Syntactic projections and the lexicon. In Moore, John & Polinsky, Maria (eds.), Explanation in linguistic theory, 31–67. Standford: CSLI.

Borer, Hagit. 2005. Structuring sense: The normal course of events, vol. 2. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199263929.001.0001

Bowers, John. 1993. The syntax of predication. Linguistic Inquiry 24(4). 591–656.

Burzio, Luigi. 1986. Italian syntax: A Government-Binding approach. Dordrecht: Reidel. DOI:  http://doi.org/10.1007/978-94-009-4522-7

Cecchetto, Carlo. 2000. Doubling structures and reconstruction. Probus 12. 93–126. DOI:  http://doi.org/10.1515/prbs.2000.12.1.93

Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.

Chomsky, Noam. 2004. Beyond explanatory adequacy. In Belletti, Adriana (ed.), Structures and beyond: The cartography of syntactic structures 3, 104–131. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780195171976.003.0004

Cifuentes Honrubia, José Luis. 1999. Bases sintácticas y bases semánticas de la inacusatividad en verbos de movimiento. Revista de investigación lingüística 2(2). 37–72.

D’Alessandro, Roberta & Franco, Irene & Gallego, Ángel J. (eds.). 2017. The verbal domain. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198767886.001.0001

Davies, Mark. on line. Corpus del español NOW corpus. Available online at https://www.corpusdelespanol.org/hist-gen/.

Doron, Edit & Hovav, Malka Rappaport. 2009. A unified approach to reflexivization in Semitic and Romance. Brill’s Journal of Afroasiatic Languages and Linguistics 1. 75–105. DOI:  http://doi.org/10.1163/187666309X12491131130503

Dowty, David. 1991. Thematic proto-roles and argument selection. Language 67(3). 574–619. DOI:  http://doi.org/10.1353/lan.1991.0021

Embick, David. 2004. On the structure of resultative participles in English. Linguistic Inquiry 35(3). 355–392. DOI:  http://doi.org/10.1162/0024389041402634

Fábregas, Antonio. 2021. SE in Spanish. Borealis – An International Journal of Spanish Linguistics 10(2). 1–236. DOI:  http://doi.org/10.7557/1.10.2.5934

Folli, Raffaella & Harley, Heidi. 2005. Flavors of v. In Kempchinsky, Paula Marie & Slabakova, Roumyana (eds.), Aspectual inquiries, 99–120. Dordrecht: Springer. DOI:  http://doi.org/10.1007/1-4020-3033-9_5

Folli, Raffaella & Harley, Heidi. 2008. Teleology and animacy in external arguments. Lingua 118. 190–202. DOI:  http://doi.org/10.1016/j.lingua.2007.02.004

Grano, Thomas. 2016. A coercion-free semantics for intend. In Proceedings of the 51st annual Chicago Linguistic Society. 213–223. Chicago IL: Chicago Linguistic Society.

Harley, Heidi. 2003. Possession and the double object construction. In Pica, Pierre & Rooryck, Johany (eds.), Linguistic variation yearbook 2, 31–70. Amsterdam: John Benjamins. DOI:  http://doi.org/10.1075/livy.2.04har

Harley, Heidi. 2005. How do verbs get their names? Denominal verbs, manner incorporation, and the ontology of verb roots in English. In Erteschik-Shir, Nomi & Rapaport, Tova (eds.), The syntax of aspect: Deriving thematic and aspectual interpretation, 42–64. New York, NY: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199280445.003.0003

Harley, Heidi. 2013. External arguments and the Mirror Principle: On the distinctness of Voice and v. Lingua 125. 34–57. DOI:  http://doi.org/10.1016/j.lingua.2012.09.010

Harley, Heidi. 2017. The “bundling” hypothesis and the disparate functions of little v. In D’Alessandro, Roberta & Franco, Irene & Gallego, Ángel J. (eds.), The verbal domain, 3–28. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198767886.003.0001

Hoekstra, Teun. 1988. Small clause results. Lingua 74(2–3). 101–139. DOI:  http://doi.org/10.1016/0024-3841(88)90056-3

Hoekstra, Teun. 1992. Small clause theory. Belgian Journal of Linguistics 7. 125–151. DOI:  http://doi.org/10.1075/bjl.7.08hoe

Homer, Vincent & Sportiche, Dominique. 2011. Clitics in French causatives: wait and see. In Paper presented at GLOW 34. University of Vienna, 28 April.

Huang, C.-T. James. 1982. Logical relations in Chinese and the theory of grammar: Massachusetts Institute of Technology dissertation.

Kratzer, Angelika. 1996. Severing the external argument from its verbs. In Zaring, Laurie Ann & Rooryck, Johan (eds.), Phrase structure and the lexicon, 109–137. Dordrecht: Kluwer. DOI:  http://doi.org/10.1007/978-94-015-8617-7_5

Levin, Beth. 1983. On the nature of ergativity. Cambridge, MA: Massachusetts Institute of Technology dissertation.

Levin, Beth & Hovav, Malka Rappaport. 1995. Unaccusativity: At the syntax - lexical semantics interface. Cambridge: MIT press.

Marantz, Alec. 1984. On the nature of grammatical relations. Cambridge, MA: MIT Press.

Marantz, Alec. 1997. No escape from syntax: Don’t try morphological analysis in the privacy of your own lexicon. In University of Pennsylvania working papers in linguistics, vol. 4, 201–225. Penn Graduate Linguistics Society.

Marantz, Alec. 2005. Objects out of the lexicon: Objects as events. Ms., Massachusetts Institute of Technology.

Marantz, Alec. 2013. Verbal argument structure: Events and participants. Lingua 130. 152–168. DOI:  http://doi.org/10.1016/j.lingua.2012.10.012

Martin, Fabienne & Schäfer, Florian & Kastner, Itamar. 2023. The lexical pragmatics of reflexive marking. [Manuscript submitted for publication].

Mateu, Jaume. 2002. Argument structure: relational construal at the syntax - semantics interface: Universitat Autònoma de Barcelona dissertation.

Mateu, Jaume & Massanell i Messalles, Mar. 2015. A constructional approach to auxiliary selection: Evidence from existential constructions. In Kailuweit, Rolf & Rosemeyer, Malte (eds.), Auxiliary selection revisited: Gradience and gradualness, 183–212. Berlin: Mouton de Gruyter. DOI:  http://doi.org/10.1515/9783110348866-007

Mateu, Jaume & Acedo-Matellán, Víctor. 2012. The Manner/Result Complementarity revisited: A syntactic approach. In Cuervo, M. Cristina & Roberge, Yves (eds.), The end of argument structure? Syntax and semantics, 209–228. New York, NY: Academic Press. DOI:  http://doi.org/10.1163/9781780523774_010

Merchant, Jason. 2013. Voice and ellipsis. Linguistic Inquiry 44(1). 77–108. DOI:  http://doi.org/10.1162/LING_a_00120

Nunberg, Geoffrey & Sag, Ivan A. & Wasow, Thomas. 1994. Idioms. Language 70(3). 491–538. DOI:  http://doi.org/10.1353/lan.1994.0007

Perlmutter, David. 1978. Impersonal passives and the unaccusative hypothesis. In Proceedings of the Fourth Annual Meeting of the Berkeley Linguistics Society. 157–189. DOI:  http://doi.org/10.3765/bls.v4i0.2198

Pineda, Anna. 2020. La (in)transitivitat en Fabra: aspectes clau. In Murgades, Josep & Nogué, Neus & Bellés, Eloi (eds.), Pompeu Fabra, a la universitat i de la universitat, 139–165. Barcelona: Edicions de la Universitat de Barcelona.

Pineda, Anna & Berro, Ane. 2020. Hybrid intransitives in Basque. Glossa: A Journal of General Linguistics 5(1). 1–28. DOI:  http://doi.org/10.5334/gjgl.824

Pylkkänen, Liina. 2008. Introducing arguments. Cambridge, MA: MIT Press. DOI:  http://doi.org/10.7551/mitpress/9780262162548.001.0001

Ramchand, Gillian. 2008. Verb meaning and the lexicon. Cambridge: Cambridge University Press. DOI:  http://doi.org/10.1017/CBO9780511486319

Rappaport Hovav, Malka. 2014. Building scalar changes. In Alexiadou, Artemis & Borer, Hagit & Schäfer, Florian (eds.), The syntax of roots and the roots of syntax, 259–281. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199665266.003.0012

Rappaport Hovav, Malka. 2020. Deconstructing internal causation. In Siegal, Elitzur Bar-Asher & Boneh, Nora (eds.), Perspectives on causation: Selected papers from the Jerusalem 2017 workshop, 219–255. Dordrecht: Springer. DOI:  http://doi.org/10.1007/978-3-030-34308-8_7

Rappaport Hovav, Malka & Levin, Beth. 1998. Building verb meanings. In Butt, Miriam & Geuder, Wilhelm (eds.), The projection of arguments: Lexical and compositional factors, 97–134. Standford, CA: CSLI Publications.

Rappaport Hovav, Malka & Levin, Beth. 2010. Reflections on Manner/Result Complementarity. In Hovav, Malka Rappaport & Doron, Edit & Sichel, Ivy (eds.), Syntax, lexical semantics, and event structure, 21–38. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199544325.003.0002

Real Academia Española, RAE. on line. Banco de datos CREA. Corpus de referencia del español actual. Available at http://corpus.rae.es/creanet.html.

Rigau, Gemma. 1990. Les propietats d’agradar: Estructura temàtica i comportament sintàctic. Caplletra. Revista Internacional de Filologia 8. 7–20.

Rigau, Gemma. 1991. On the functional properties of AGR. Catalan Working Papers in Linguistics 1. 235–260.

Rigau, Gemma. 1992. Aspects of Catalan dialect syntax. Unaccusative sentences and related structures. Journal of Hispanic Research 2(2). 315–329.

Rigau, Gemma. 1994. Les propietats dels verbs pronominals. Els Marges 50. 29–39.

Rigau, Gemma. 1997a. Locative sentences and related constructions in Catalan: ésser / haver alternation. In Mendikoetexea, Amaya & Uribe-Etxebarria, Myriam (eds.), Theoretical issues at the morphology-syntax interface, 395–421. Euskal Herriko Unibertsitatea / Gipuzkoa Foru Aldundia.

Rigau, Gemma. 1997b. Sintaxi i variació dialectal: Les construccions locatives. In Lloret, Maria-Rosa & Boix, Emili & Lorente, Mercè & Payrató, Lluís & Perea, Maria-Pilar (eds.), Anàlisi de la variació lingüística, 179–201. Barcelona: Promocions i Publicacions Universitàries, Universitat de Barcelona.

Rooryck, Johan & Wyngaerd, Guido Vanden. 2011. Dissolving binding theory. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/acprof:oso/9780199691326.001.0001

Ross, John Robert. 1967. Constraints on variables in syntax: Massachusetts Institute of Technology dissertation.

Rosselló, Joana. 2008. El SV, I: Verbs i arguments verbals. In Solà, Joan & Lloret, Maria-Rosa & Mascaró, Joan (eds.), Gramàtica del català contemporani, 1853–1949. Barcelona: Empúries.

Schäfer, Florian. 2008. The causative alternation. Language and Linguistics Compass 3(2). 641–681. DOI:  http://doi.org/10.1111/j.1749-818X.2009.00127.x

Schäfer, Florian. 2017. Romance and Greek medio-passives and the typology of Voice. In D’Alessandro, Roberta & Franco, Irene & Gallego, Ángel J. (eds.), The verbal domain, 129–152. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198767886.003.0006

Solà, Joan. 1973. Estudis de sintaxi catalana. Barcelona: Edicions 62 (Llibres a l’Abast, 105).

Solà, Joan. 1987. Qüestions controvertides de sintaxi catalana. Barcelona: Edicions 62 (Llibres a l’Abast, 229).

Solà, Joan. 1994. Sintaxi normativa: Estat de la qüestió. Barcelona: Empúries.

Solà, Joan. 1999. Parlem-ne. Converses lingüístiques. Palma: Moll.

Solstad, Torgrim & Bott, Oliver. 2017. Causality and causal reasoning in natural language. In Waldmann, Michael R. (ed.), The Oxford handbook of causal reasoning, 619–644. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oxfordhb/9780199399550.013.32

Sorace, Antonella. 2000. Gradients in auxiliary selection with intransitive verbs. Language 859–890. DOI:  http://doi.org/10.2307/417202

Sportiche, Dominique. 2014. Assessing unaccusativity and reflexivity: Using focus alternatives to decide what gets which θ-role. Linguistic Inquiry 45. 305–321. DOI:  http://doi.org/10.1162/LING_a_00156

Sportiche, Dominique. 2020. Constraints on reflexivization. Ms., University of California, Los Angeles.

Sundaresan, Sandhya & McFadden, Thomas. 2017. The articulated v layer: Evidence from Tamil. In D’Alessandro, Roberta & Franco, Irene & Gallego, Ángel J. (eds.), The verbal domain, 153–178. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198767886.003.0007

Talmy, Leonard. 1975. Semantics and syntax of motion. In Kimball, John P. (ed.), Syntax and semantics, vol. 4, 181–238. New York, NY: Academic Press. DOI:  http://doi.org/10.1163/9789004368828_008

Tollan, Rebecca. 2018. Unergatives are different: Two types of transitivity in Samoan. Glossa: A Journal of General Linguistics 3(35). 1–41. DOI:  http://doi.org/10.5334/gjgl.223

Tollan, Rebecca & Oxford, Will. 2018. Voice-less unergatives: Evidence from Algonquian. In Bennett, Wm. G. & Hracs, Lindsay & Storoshenko, Dennis Ryan (eds.), Proceedings of the 35th West Coast Conference on Formal Linguistics, 399–408. Cascadilla Proceedings Project.

Torrego, Esther. 1994. On the nature of clitic doubling. In Campos, Héctor & Kempchinsky, Paula (eds.), Evolution and revolution in linguistic theory, 399–418. Washington DC: Georgetown University Press.

Tubino-Blanco, Mercedes. 2020. La involuntariedad: ¿ambigüedad sintáctica o vaguedad referencial? In García, Fernando López (ed.), La involuntariedad en español, 79–104. Berlin: Peter Lang.

Uriagereka, Juan. 1988. On government: University of Connecticut dissertation.

Uriagereka, Juan. 2001. Doubling and possession. In Gerlach, Birgit & Grijzenhout, Janet (eds.), Clitics in phonology, morphology and syntax, 405–431. Amsterdam: John Benjamins. DOI:  http://doi.org/10.1075/la.36.16uri

Van Valin, Robert & Wilkins, David. 1996. The case for ‘effector’: Case roles, agents, and agency revisited. In Shibatani, Masayoshi & Thompson, Sandra (eds.), Grammatical constructions: Their form and meaning, 289–322. Oxford: Oxford University Press. DOI:  http://doi.org/10.1093/oso/9780198235392.003.0011

von Stechow, Arnim. 1996. The different readings of wieder ‘again’: A structural account. Journal of Semantics 13(2). 87–138. DOI:  http://doi.org/10.1093/jos/13.2.87

Wechsler, Stephen. 2005. What is right and wrong about little v. In Dimitrova-Vulchanova, Mila & Åfarli, Tor A. (eds.), Grammar and beyond—Essays in honour of Lars Hellan. 179–195. Oslo: Novus Press.

Wood, Jim & Marantz, Alec. 2017. The interpretation of external arguments. In D’Alessandro, Roberta & Franco, Irene & Gallego, Ángel J. (eds.), The verbal domain, 255–278. Oxford: Oxford University Press.