1. Introduction

Complex predicates are observed cross-linguistically and have been widely discussed in the literature, in particular those involving the combination of a light verb with a deverbal noun (Jespersen 1965, p. 117). The following examples illustrate light verb combinations: (1a) and (2a) show the monotransitive verb ‘to give’ with deverbal nouns deriving from unergative verbs, in both English and European Portuguese (EP); (1b-d) and (2b-d) show the ditransitive ’to give’, which selects a direct object (DO), typically an indefinite DP, and an indirect object (IO) [±ANIM].

    1. (1)
    1. (a)
    1. to give a scream / a smile / a laugh
    1.  
    1. (b)
    1. to give someone advice / information / (emotional) support
    1.  
    1. (c)
    1. to give someone a kiss / a push / a hug
    1.  
    1. (d)
    1. to give the car a wash, to give the soup a stir
    2. (adapted from Butt & Geuder, 2001, pp. 339–340)
    1. (2)
    1. (a)
    1. dar
    2. give
    1. um
    2. a
    1. grito /
    2. scream /
    1. um
    2. a
    1. passeio
    2. walk’
    1. ‘to scream / to take a walk’
    1.  
    1. (b)
    1. dar
    2. give
    1. um
    2. an
    1. conselho /
    2. advice /
    1. uma
    2. a
    1. ajuda /
    2. help /
    1. apoio
    2. support
    1. a
    2. to
    1. alguém
    2. someone
    1. ‘to advice / to help / to support someone’
    1.  
    1. (c)
    1. dar
    2. give
    1. um
    2. a
    1. abraço /
    2. hug /
    1. um
    2. a
    1. empurrão /
    2. push /
    1. um
    2. a
    1. beijo
    2. kiss
    1. a
    2. to
    1. alguém
    2. someone
    1. ‘to hug / push / kiss someone’
    1.  
    1. (d)
    1. dar
    2. give
    1. uma
    2. a
    1. pintura
    2. painting
    1. à
    2. to.the
    1. casa /
    2. house /
    1. dar
    2. give
    1. uma
    2. a
    1. leitura
    2. reading
    1. ao
    2. to.the
    1. artigo
    2. paper
    1. ‘to paint the house / read the paper’
    2. (adapted from Gonçalves et al., 2010, pp. 451, 458)

Butt and Geuder (2001) discuss an array of different uses of light verb ‘to give’ in English, which they characterize as “continual extensions from the prototypical concrete meaning of give, with the extension of the meaning being correlated with the increasing abstractness of the referents filling the Theme argument slot” (p. 343) (see also Newman 1996). For instance, they consider that monotransitive instances of give, as illustrated in (1a) and (2a), only convey the notion of emission, differently from ditransitive uses of give, in which there is proper transfer of possession, which sometimes can also imply transfer in space, as in give him the ball or give the dog a bone.1 Furthermore, there are combinations, such as (1b) and (2b), that neither imply transfer of possession nor transfer in space of a concrete entity; at this point, Butt and Geuder (2001) distinguish between combinations as to give someone advice from to give someone emotional support, arguing that in the former there is still a controller of the event, while in the latter the combination “only describe[s] a beneficial effect for the Recipient, but drop[s] the notion of control over these effects altogether” (p. 341). Finally, combinations such as (1c-d) an (2c-d) are described as a mere exertion of some effect on the Recipient, which can also apply to inanimate concrete objects (see (1d) and (2d)). Butt and Geuder (2001, p. 342) refer that, in examples (1d) and (2d), “the mere application of an action to an entity is all that is left of the transaction meaning of give.” These distinct uses of light ‘to give’ are reminiscent of Huddleston and Pullum’s (2002) syntactic-semantic proposal of different subclasses of the light verb ‘to give’.

In this paper, we aim to extend the discussion of the increment of “lightness” of ‘to give’ to Portuguese data. First, we will discuss the possibility that nominal expressions co-occurring with the light verb dar are not exclusively event-denoting. At this point, it is crucial to recall that the relevant notion of event used by Butt & Geuder (2001), among others, is the one presented by Grimshaw (1990), who shows that some deverbal nouns may have two readings, namely an eventive reading, with a certain degree of abstractness, and a resultative reading (of the event), with a certain degree of concreteness. Moreover, some authors emphasized that, besides the distinction between event and result, the notions of result and entity must be distinguished (e.g., Brito & Oliveira 1997; Sleeman & Brito 2010). The result may still be the final phase of an accomplishment/culminated process; on the contrary, an entity is the material or concrete result/product of some event. Bearing in mind this distinction, Choupina and Brito (2018) argued that some Ns that combine with the light verb dar in Portuguese, such as dar uma ajuda lit. ‘to give a help’ or dar um empurrão lit. ‘to give a push’ (see (1b-c) and (2b-c)), can have hybrid interpretations between the entity or the eventive reading. Note that they can be affected by numerals and quantifiers, as in dar várias ajudas lit. ‘to give several helps’ or dar dois empurrões lit. ‘to give two pushes’.2 Therefore, one may ask whether nominal expressions used in complex predicates with light verb dar may denote events and/or results, exhibiting different degrees of abstractness/concreteness. We will address and develop this idea in section 3.

We will also argue that there is an additional use of dar in which only a causative interpretation (not a transference one), is available, as illustrated in English, Spanish, and Portuguese examples in, respectively, (3), (4) and (5–6).

    1. (3)
    1. (a)
    1. That gave us a laugh! (laugher = us)
    1.  
    1. (b)
    1. That gave me a shudder! (shudderer = me)
    2. (Bruening, 2015, p. 19)
    1. (4)
    1. Al
    2. the
    1. técnico
    2. coach.dat
    1. le
    2. Cl.DAT
    1. dan
    2. give
    1. rabia
    2. fury
    1. las
    2. the
    1. protestas.
    2. complaints
    1. ‘The complaints make the coach furious.’
    2. (adapted from Cuervo, 2008, p. 140)
    1. (5)
    1. O
    2. the
    1. João
    2. João
    1. deu
    2. gave
    1. preocupações
    2. worries
    1. ao
    2. to.the
    1. pai.
    2. father
    1. ‘John worried his father.’
    2. (adapted from Duarte et al., 2009, p. 5)
    1. (6)
    1. (a)
    1. Elogios
    2. compliments
    1. lhe
    2. 3.SG.DAT
    1. dão
    2. gave
    1. alegria.
    2. joy
    1. ‘Compliments give him/her joy.’
    1.  
    1. (b)
    1. Comer
    2. eat
    1. dá
    2. give
    1. pressão
    2. pressure
    1. alta.
    2. high
    1. ‘Eating gives high pressure.’
    2. (adapted from Rassi & Vale, 2013, p. 116)

Both Cuervo (2008), for Spanish, and Duarte et al. (2009), for Portuguese, argue that dar can combine with psychological Ns, yielding causative structures. However, examples (3–6) do not exclusively involve psychological Ns, but also Ns denoting illnesses (cf. (6b)), as pointed out by Rassi & Vale (2013).3 Notice also that examples such as (5–6) suggest that dar combines not only with deverbal nominalizations denoting events (independently of their abstractness/concreteness), but also with Ns that are not morphologically related to Vs, and denote states, as the deadjectival N alegria ‘joy’ and the underived N pressão ‘pressure’. Moreover, Cuervo (2008) argues that, when dar occurs with psychological Ns, the syntactic structure is different from complex predicates with a derived N, i.e., a deverbal N.4 In her approach, predicates with dar + psychological Ns are unaccusative structures. We will argue against this proposal in section 4.

This brief presentation shows that the use of light verb ‘to give’ is not uniform; there appear to be different subclasses with regard to their syntactic and semantic properties. In fact, Butt’s (2010, pp. 52–53) states that “complex predicate may range over different types and therefore exhibit differing syntactic and semantic properties. As a consequence (…), light verbs constitute a cohesive class on the one hand but fall into differing subclasses on the other.”

Therefore, the main goals of this paper are a) to discuss whether there are different subclasses of complex predicates with the light verb dar in African Varieties of Portuguese (AVPs), using spoken urban corpora; b) to discuss to what extent AVPs contribute to a better understanding of complex predicates in general; and, finally, c) whether different subclasses of complex predicates have distinct underlying syntactic representations.

The paper is organized as follows. Section 2 describes the methodology. Section 3 provides a corpus-based analysis of the complex predicates with the light verb dar in the three above-mentioned AVPs. In particular, we will take both morphological and syntactic-semantic properties as cues to draw a classification of complex predicates with dar. Section 4 discusses the underlying syntactic structures of transferential and causative complex predicates with light verb dar. We will mainly focus on the argumental status of IOs, their thematic role, and the semantic features that distinguish the predicates. Section 5 summarizes the findings of this study.

2. Methodology

Our case study of complex predicates is based on comparable spoken urban corpora of Mozambican Portuguese (MOP), Angolan Portuguese (AP), and Santomean Portuguese (STP), that were prepared within the project Possession and Location: microvariation in African varieties of Portuguese (PALMA). The corpora were collected in capitals Maputo, Luanda, and São Tomé and Príncipe, between 2008 and 2020, by researchers of the Center of Linguistics of the University of Lisbon, as well as doctoral students, through semi-structured interviews to 205 informants. The corpus consists of 108 hours of recording, corresponding to over one million tokens (1,097,702), and is fairly well balanced with respect to the main sociolinguistic variables (gender, age, and schooling).

It was annotated with part-of-speech (POS) and lemma information, and it has been made searchable on the CQPweb platform (see Hagemeijer et al. 2022 for further details).

To achieve our goals, a particular data set was organized. We extracted all the occurrences of the verb dar from the three corpora by using lemma and POS tags at the query node, namely {dar}_V*. Then, we proceeded to exclude the following contexts: (i) dar as a core dative verb of transfer of possession, which was already extensively discussed in previous literature on AVPs (see Brito 2008, 2022; Gonçalves, Duarte & Hagemeijer 2022, a.o.); (ii) idiomatic expressions and collocations involving the verb dar (e.g., dar cabo de lit. ‘give away with’, dar sinal lit ‘give signal’, dar espaço lit. ‘give space’, dar licença lit. ‘to give permission’, dar uma mão lit. ‘to give a hand’, etc.); (iii) impersonal uses of dar (e.g., dar-se a comunicação lit. ‘to give the communication’, dar-se o caso lit. ‘to give the case’, etc.); (iv) unclear contexts, due to hesitations and reformulations, as well as cases that were ambiguous, difficult to categorize as a complex predicate.

The remaining data encompass 407 complex predicates that were organized according to the following variables: (i) the existence of a full-fledged verb able to paraphrase the complex predicate; (ii) the typology of the N involved in the complex predicate: Ns morphologically related to Vs, ADJs, and Ns, including deverbal, deadjectival, and denominal Ns (and Ns that give rise to parasynthetic verbs), as well as underived Ns; (iii) the main reading of the N (event, event-result, and state, as in psychological/physical sensations); and (iv) the form of the DP Theme, in particular, the determination (bare noun, definite, indefinite) and its singular or plural form; (v) the type of ditransitive structures involved: Ditransitive Prepositional Construction (DPC), with specification of the preposition used, or Double Object Construction (DOC).

3. Complex predicates with the light verb dar ‘to give’

3.1. General data

Complex predicates with dar are productive in the corpus we analyzed (407 occurrences), notwithstanding slight differences between AVPs. Table 1 summarizes the distribution of data by morphological type of N and AVP.

Table 1: Occurrences of light verb dar by AVP and morphological type of the N.

AP MOP STP TOTAL
N morphologically related to V 96 87 127 310 (76,2%)
N morphologically related to ADJ 9 11 11 31 (7,6%)
N morphologically related to N 5 3 3 11 (2,7%)
Underived N 11 24 20 55 (13,5%)
TOTAL 121 125 161 407

As expected, complex predicates with the light verb dar in our corpus involve a deverbal noun or are morphologically related to a V (76,2%), but they are not exclusive. We also found some combinations involving Ns morphologically related to ADJs and Ns, although on a limited scale (7,6% and 2,7% respectively). Moreover, 13,5% of the occurrences correspond to complex predicates which involve underived Ns, but which are in some way predicative, as we will argue below. It is also worth mentioning that 88,2% of the complex predicates we analyze can be paraphrased by a main verb with the semantic reading of the combination light verb + N, irrespective of the morphological derivation of the N. Therefore, 11,8% of the complex predicates under analysis do not have a corresponding verb.

In the next sections, we will present data from the three AVP, by morphological type of N, in order to discuss the internal structure of the DP, the semantic reading of the combination with the light verb, and the type of light verb dar.

3.1.1 Complex predicates with Ns morphologically related to Vs

Among the 310 occurrences in the corpus of light verb dar + Ns morphologically related to a V5, we found 116 different combinations. Most of them have a corresponding transitive main verb (258/310, i.e., 83,2%), as in (7a) and (8–9), respectively dar emprego lit. ‘to give employment’ / empregar ‘to employ’, dar educação lit. ‘to give education’ / educar ‘to educate’, dar indemnização lit. ‘to give compensation’ / indemnizar ‘to compensate’, dar encorajamento lit. ‘to give encouragement’ / encorajar ‘to encourage’, dar fervura lit. ‘to give boil’ / ferver ‘to boil’. But we also found a few instances in the corpus which involve Ns that are derived from unergative verbs, such as dar um grito lit. ‘to give a scream’ / gritar ‘to scream’ (10/310, i.e., 3,2%) (cf. (7b)). In our analysis, these combinations of dar + event N bear an event reading.

    1. (7)
    1. (a)
    1. estamos
    2. are
    1. a
    2. to
    1. dar
    2. give
    1. emprego
    2. employment
    1. a
    2. to
    1. tanta
    2. so many
    1. gente (AP)
    2. people
    1. ‘We employ so many people.’
    1.  
    1. (b)
    1. eu
    2. I
    1. dei
    2. gave
    1. um
    2. a
    1. grito (AP)
    2. scream
    1. ‘I screamed.’
    1. (8)
    1. (a)
    1. meu
    2. my
    1. sonho
    2. dream
    1. seria
    2. would be
    1. dar
    2. give
    1. educação
    2. education
    1. aos
    2. to.the
    1. meus
    2. my
    1. filhos (MOP)
    2. children
    1. ‘My dream would be to give education to my children’.
    1.  
    1. (b)
    1. trabalhava
    2. worked
    1. na
    2. in.the
    1. África do Sul,
    2. South Africa,
    1. deram
    2. gave
    1. uma
    2. a
    1. indemnização (MOP)
    2. compensation
    1. ‘I worked in South Africa, they gave [me] compensation.’
    1. (9)
    1. (a)
    1. essa
    2. that
    1. galinha
    2. chicken
    1. que
    2. that
    1. deu
    2. gave
    1. ela
    2. she
    1. encorajamento (STP)
    2. encouragement
    1. ‘That chicken that gave her encouragement…’
    1.  
    1. (b)
    1. deixa-se
    2. let-SE
    1. dar
    2. give
    1. umas
    2. some
    1. ferveduras (STP)
    2. boils
    1. ‘Let it boil.’

In what concerns the internal structure of the Theme DP, we have found many occurrences of bare nouns (125/310 occurrences, i.e., 40,3%) (see (7a–9a)), which are almost exclusively singular, along with Ns preceded by indefinite articles (99/310 occurrences, i.e., 31,9%) (see (7b–9b)). These occurrences correspond to 72,3% of the total analyzed with Ns morphological related to Vs. The use of definite articles, possessive pronouns, demonstratives, numerals, and quantifiers, although available, has low frequencies in complex predicates with the light verb dar in the corpus (86/310 occurrences, i.e., 27,7%) (see (10)).

    1. (10)
    1. (a)
    1. estão
    2. are
    1. a
    2. to
    1. dar
    2. give
    1. o
    2. the
    1. abate
    2. cut
    1. em
    2. in
    1. muitas
    2. many
    1. árvores. (AP)
    2. trees
    1. ‘They are cutting down many trees.’
    1.  
    1. (b)
    1. ele
    2. he
    1. dá-me
    2. gives-1SG.DAT
    1. todo
    2. all
    1. o
    2. the
    1. apoio (MOP)
    2. support
    1. ‘He gives me all the support.’
    1.  
    1. (c)
    1. deu-lhe
    2. gave-3SG.DAT
    1. dois
    2. two
    1. soco (STP)
    2. punch
    1. ‘He/she gave him / her two punches.’

As mentioned in section 1, according to Grimshaw (1990), Ns morphologically related to Vs have two readings: the event reading and the result reading. In fact, data from the AVPs show that Ns morphologically related to Vs that combine with the light verb dar do not always have an exclusive event reading, as in (7–10). The corpus also exhibits combinations of dar with Ns morphologically related to Vs that can show a double semantic interpretation of event-result (11), and denote states (12).

    1. (11)
    1. (a)
    1. para
    2. to
    1. dar
    2. give
    1. alimentação
    2. food
    1. aos
    2. to.the
    1. miúdos (AP)
    2. kids
    1. ‘to feed the kids.’
    1.  
    1. (b)
    1. para
    2. to
    1. dar
    2. give
    1. a
    2. the
    1. medicação
    2. medication
    1. aos
    2. to.the
    1. doentes (MOP)
    2. patients
    1. ‘to give patients medication.’
    1.  
    1. (c)
    1. para
    2. to
    1. poder
    2. be able
    1. dar
    2. give
    1. o
    2. the
    1. sustento
    2. sustenance
    1. às
    2. to.the
    1. suas
    2. their
    1. famílias (MOP)
    2. families
    1. ‘to be able to sustenance their families.’
    1. (12)
    1. (a)
    1. Ele
    2. He
    1. está
    2. is
    1. me
    2. 3SG.DAT
    1. dar
    2. give
    1. muito
    2. a lot of
    1. ânimo. (STP)
    2. encouragement
    1. ‘He is strongly encouraging me.’
    1.  
    1. (b)
    1. o
    2. the
    1. álcool
    2. alcohol
    1. dá
    2. gives
    1. coragem
    2. courage
    1. para
    2. to
    1. fazer
    2. do
    1. bandidagem (AP)
    2. criminality
    1. ‘Alcohol encourages criminal behavior.’
    1.  
    1. (c)
    1. as
    2. the
    1. coisas
    2. things
    1. negativas
    2. negative
    1. que
    2. that
    1. acontecem
    2. happen
    1. com
    2. with
    1. a
    2. the
    1. gente
    2. people
    1. nos
    2. 1PL.DAT
    1. dão
    2. give
    1. motivação (MOP)
    2. motivation
    1. ‘The negative things that happen to us give us motivation.’

In fact, combinations such as (11) may be interpreted as alimentar ‘to feed’, medicar ‘to medicate’, sustentar ‘to sustain’ and, in this case, they favor the event interpretation, but they also have the meaning of ’to give (some) food’, to give (some) medication’, ‘to give (some) sustenance’, which favor the result reading. Moreover, Portuguese has “short” alternatives to some of these nouns, the so-called post-verbal nouns / “regressive” nouns, formed by conversion (see alimento vs. alimentação ‘food’), which contributes to the view that these Ns indeed may have in the corpus an event/result reading. Moreover, there are also predicates such as dar ânimo lit. ‘to give encouragement’, dar motivação lit. ‘to give motivation’, dar conforto lit. ‘to give comfort’, in which dar combines with psychological state Ns: the IO is modified by the action expressed by the V. Therefore, we must conclude that, when dar combines with Ns morphologically related to a V in our data, the complex predicate can either have an exclusive event reading or, instead, a double event-result reading.

Data corresponding to complex predicates with Ns morphologically related to Vs but whose meaning cannot be obtained from it (42/310, i.e., 13,5%) are illustrated in (13), respectively dar queda lit. ‘to give drop’, dar medo lit ‘to give fear’, and dar vómitos lit. ‘to give vomits’.

    1. (13)
    1. (a)
    1. a
    2. the
    1. gente
    2. people
    1. vai
    2. go
    1. bebendo,
    2. drinking,
    1. aquilo
    2. that
    1. é
    2. is
    1. doce,
    2. sweet,
    1. depois
    2. then
    1. dá-nos
    2. gives-1PL.DAT
    1. uma
    2. a
    1. queda (AP)
    2. drop
    1. ‘While we are drinking it, it tastes sweet, then it gives us a hangover.’
    1.  
    1. (b)
    1. dava
    2. gave
    1. medo
    2. fear
    1. pessoa
    2. person
    1. sair
    2. leave
    1. para
    2. to
    1. ir
    2. go out
    1. na
    2. in.the
    1. discoteca (AP)
    2. disco
    1. ‘One would be(come) afraid to go to the disco.’
    1.  
    1. (c)
    1. paludismo
    2. malaria
    1. é
    2. is
    1. a
    2. the
    1. febre (…)
    2. fever (…)
    1. há
    2. there are
    1. quem
    2. who
    1. dá
    2. gives
    1. vómitos (STP)
    2. vomits
    1. ‘Malaria is the fever. There are people who throw up.’

In fact, although the verbs cair ‘to fall’, amedrontar ‘to frighten’, and vomitar ‘to throw up’ do exist in Portuguese, they cannot be used as corresponding transitive main verbs to the combinations presented above. Therefore, we will argue that, in structures such as (13), which do not encode a transference meaning, dar bears a causative meaning, an issue to which we will return to in the next section.

3.1.2 Complex predicates with Ns morphologically related to ADJs and Ns

In Table 1 above, it was shown that the corpus exhibits 10,3% (42/407) of occurrences of light verb dar + Ns morphologically related to ADJs and Ns (respectively 7,6% and 2,7%), corresponding to a total of 16 different combinations. Similarly to most Ns we discussed so far, the majority of the complex predicates with denominal and deadjectival Ns we analyzed can be paraphrased by a main verb. See (14), for complex predicates with deadjetival Ns, respectively dar prioridade lit. ‘to give priority’ / priorizar ‘to prioritise’, dar autonomia lit. ‘to give autonomy’ / autonomizar ‘to autonomise’, and dar continuidade lit. ‘to give continuity’ / continuar ‘to continue’. For complex predicates with denominal Ns, consider the examples presented in (15), respectively dar uma bofetada lit. ‘to give a slap’ / esbofetear ‘to slap’, dar uma cabeçada ‘to give a headbutt’ / cabecear ‘to headbutt’6, and dar uma chicotada lit. ‘to give a whip’ / chicotear ‘to whip’.

    1. (14)
    1. (a)
    1. os
    2. the
    1. adultos
    2. adults
    1. deram
    2. gave
    1. prioridade
    2. priority
    1. a
    2. to
    1. isso (AP)
    2. this
    1. ‘Adults prioritised this.’
    1.  
    1. (b)
    1. darem
    2. give
    1. uma
    2. an
    1. autonomia
    2. autonomy
    1. à
    2. to.the
    1. língua
    2. language
    1. moçambicana (MOP)
    2. mozambican
    1. ‘Empower the Mozambican language’
    1.  
    1. (c)
    1. vou
    2. go
    1. dar
    2. give
    1. continuidade
    2. continuity
    1. aos
    2. to.the
    1. meus
    2. my
    1. estudos (STP)
    2. studies
    1. ‘I will continue my studies.’
    1. (15)
    1. (a)
    1. deram-lhe
    2. gave-3SG.DAT
    1. duas
    2. two
    1. bofetadas (AP)
    2. slaps
    1. ‘They slapped him/her twice.’
    1.  
    1. (b)
    1. as
    2. the
    1. minhas
    2. my
    1. amigas
    2. friends
    1. me
    2. 1SG.DAT
    1. davam
    2. gave
    1. cabeçada (MOP)
    2. headbutt
    1. ‘My friends would headbutt me’
    1.  
    1. (c)
    1. o
    2. the
    1. pai
    2. father
    1. deu-lhe
    2. gave-3SG.DAT
    1. uma
    2. a
    1. chicotada (STP)
    2. whipping
    1. ‘The father gave him a whipping.’

Rio-Torto et al. (2013, p. 124) show that Portuguese deadjectival nouns denoting properties and states are typically abstract nouns; they occur in the singular, and they do not admit pluralization nor qualification. Indeed, deadjectival Ns in the corpus, denoting properties, and states, occur almost exclusively in the singular and are mainly formed with the suffix -idade.

Denominal nouns represent an heterogenous class, with several readings: “event, individual, local, cause/source, result, product), [that] partially coincide with the ones that characterize deverbal nouns (…)” (Rio-Torto et al., 2013, p. 131, our translation). As illustrated in (15), denominal roots in our corpus are exclusively hit and impact Ns, which are formed with the suffix -ada, which has the meaning of an event realized with the body-part or the object that the root designates (Rio-Torto et al. 2013, p. 146), respectively, cabeça ‘head’, chicote ‘whip’ and batuque ‘drumming’.7 Therefore, the Theme of the complex predicate involves a noun which, again, has a reading with a certain degree of concreteness.8

The internal structure of the Theme DP is mainly that of a bare noun (21/42 occurrences) or a noun with an indefinite article (15/42 occurrences) (cf. (16a), (16c) vs. (16b)). These occurrences correspond to 85,7% of the total analyzed with Ns morphologically related to ADJs and Ns. The use of definite articles, demonstratives, numerals, and quantifiers with these types of N in complex predicates with dar is uncommon in our data (6/42 occurrences) (see (15a) and (16)).

    1. (16)
    1. (a)
    1. vimos
    2. saw
    1. senhores
    2. men
    1. a
    2. to
    1. darem
    2. give
    1. aquelas
    2. those
    1. batucadas (AP)
    2. drums
    1. ‘We have seen men playing on the drums.’
    1.  
    1. (b)
    1. a
    2. the
    1. rádio dava
    2. radio gave
    1. a
    2. the
    1. possibilidade
    2. possibility
    1. direta
    2. direct
    1. de… (MOP)
    2. of
    1. ‘The radio gave the direct possibility of …’

Predicates involving Ns that are morphologically related to ADJs are not uniform in their semantic reading. While it is true that they mainly encode an event reading (see (14–15)), it is also true that the corpus exhibits structures as those exhibited in (17) below, in which Ns encode a stative meaning. In these cases, they can be paraphrased by a main verb: dar bebedeira lit. ‘to give drunkenness’ / embebedar-se ‘to get drunk’, dar tristeza lit. ‘to give sadness’ / entristecer ‘to sadden’, and dar fraqueza lit. ‘to give weakness’ / enfraquecer ‘to weaken’.

    1. (17)
    1. (a)
    1. dizem
    2. say
    1. que
    2. that
    1. isso
    2. this
    1. fermentar
    2. ferment
    1. é
    2. is
    1. que
    2. that
    1. dá
    2. give
    1. a
    2. the
    1. bebedeira (AP)
    2. drunk
    1. ‘They say that its fermentation is what makes people drunk.’
    1.  
    1. (b)
    1. você
    2. you
    1. vê
    2. see
    1. só
    2. just
    1. ele
    2. he
    1. a
    2. to
    1. chorar,
    2. cry,
    1. dá
    2. give
    1. uma
    2. a
    1. tristeza (AP)
    2. sadness
    1. ‘You just see him crying, it makes you sad.’
    1.  
    1. (c)
    1. paludismo
    2. malaria
    1. é
    2. is
    1. a
    2. the
    1. febre, dá
    2. fever, gives
    1. fraqueza (STP)
    2. weakness
    1. ‘Malaria is a fever, it makes you weak.’

Moreover, while one can argue that predicates of the type illustrated in (7–10) and (14–16) encode a transference meaning, independently of the degree of abstractness/concreteness, this is not true when we aim to characterize predicates such as (13) and (17). In fact, these predicates involve Ns of psychological/physical sensations, whose meaning are reminiscent of causative structures.

3.1.3 Complex predicates with underived Ns

The 55 occurrences in the corpus of complex predicates with dar and underived Ns correspond to 15 different combinations, such as dar fome lit. ‘to give hunger’, dar indigestão lit. ‘to give indigestion’, and dar dores lit. ‘to give pain’, illustrated below.

    1. (18)
    1. (a)
    1. o
    2. the
    1. retrovirais
    2. retrovirals
    1. dá
    2. give
    1. fome (AP)
    2. hunger
    1. ‘Retrovirals make you hungry.’
    1.  
    1. (b)
    1. é
    2. is
    1. por isso
    2. why
    1. que
    2. that
    1. há vez
    2. some time
    1. dá
    2. give
    1. indigestão (AP)
    2. indigestion
    1. ‘That’s why there are times it gives you indigestion.’
    1.  
    1. (c)
    1. direito
    2. law
    1. está-me
    2. is-1SG.DAT
    1. dar
    2. give
    1. dores de cabeça (STP)
    2. headaches
    1. ‘[Studying] law is giving me a headache.’

An immediate conclusion that can be drawn from the complex predicates with underived Ns, in comparison with those discussed so far, is that they typically do not have a main verb that can paraphrase its meaning. Although Vs such as esfomear ‘to starve’, and doer ‘to hurt’ exist in Portuguese, they do not capture the meaning of the examples presented in (18a) and (18c), for instance. Moreover, there is no main verb corresponding to the complex predicate in (18b).

Note that these structures involving Ns of physical sensations, which mainly occur as bare nouns (33/55 occurrences, i.e., 54,5%), are like the ones in (13) and (17), showing that causative structures with dar can occur independently of the morphological category of the N, i.e., we found evidence that they can occur with both derived and underived Ns.

As we have been showing, bare nouns are very common in the corpus we analyzed. However, this is a property of complex predicates in general, and cannot be taken as a specific property of DPs in AVPs. In section 3.2., however, we will focus on properties of complex predicates in AVPs that are not convergent with EP grammar, arguing that the observed microvariation can be explained through the predicates’ semantic features.

3.1.4. Distinguishing complex predicates beyond morphology

Butt (2010, p. 49) states that “the term complex predicate is used to designate a construction that involves two or more predicational elements (such as nouns, verbs, and adjectives) which predicate as a single unit, i.e., their arguments map onto a monoclausal syntactic structure.” The analysis developed so far lead us to conclude that the predicative Ns that form light verb combinations with dar, i.e., complex predicates, are much more diversified than what is generally assumed in the literature.9

It is widely discussed in the literature that light verb dar can combine with Ns morphologically related to Vs, which are event nouns in the sense of Grimshaw (1990) (e.g., dar emprego lit. ‘to give employment’, dar educação lit. ‘to give education’, dar encorajamento lit. ‘to give encouragement’). However, as shown above, it can also combine with denominal Ns, meaning an event realized with the body-part or the object that the root designates (e.g., dar uma bofetada lit ‘to give a slap’, dar uma cabeçada lit. ‘to give a headbutt’, dar uma chicotada lit. ‘to give a whip’), as well as with deadjectival Ns, acquiring a reading close to what Duarte & Batoréo (2004) classify as a “change of state” (e.g., dar prioridade lit. ‘to give priority’, dar autonomia lit. ‘to give autonomy’, dar continuidade lit. ‘to give continuity’). Further, as we have seen, dar also has a causative reading, irrespectively of whether the N is related to a V, an ADJ, or an underived N, as in dar medo lit ‘to give fear’, dar fraqueza lit. ‘to give weakness’, and dar fome lit. ‘to give hunger’. Next, we will discuss that the different types of combinations allowed with the light verb dar can be better understood if one uses passivization as a syntactic test to understand the syntactic and semantic features of the complex predicate.10

On the one hand, we found that there are predicates whose meaning may be obtained from an eventive passive, formed with the verb ser ‘to be’ + the past participle of the V that paraphrases the predicate. Note that the eventive passive is allowed independently of whether the N that occurs in the complex predicate is morphologically related to V (as in combinations such as dar emprego lit. ‘to give employment’, dar educação lit. ‘to give education’; dar encorajamento lit. ‘to give encouragement’), an ADJ (as in combinations such as dar prioridade lit. ‘to give priority’, dar autonomia lit. ‘to give autonomy’, dar continuidade lit. ‘to give continuity’) or a N (as in combinations such as dar uma bofetada lit. ‘to give a slap’, dar uma cabeçada ‘to give a headbutt’, and dar uma chicotada lit. ‘to give a whip’). We will classify these sorts of combinations as instances of transferential event complex predicates.

    1. (19)
    1. (a)
    1. ele foi empregado, educado, encorajado
    2. he was employed, educated, encouraged
    1.  
    1. (b)
    1. ele foi priorizado, autonomizado, continuado
    2. he was prioritised, ring-fenced, continued
    1.  
    1. (c)
    1. ele foi esbofeteado, cabeceado, chicoteado
    2. he was slapped, headbutted, whipped

On the other hand, we have also seen that dar combines with Ns with a double semantic interpretation, a hybrid reading between an event and a result, sometimes even almost a concrete entity (e.g., dar alimentação lit. ‘to give food’, dar medicação lit. ‘to give a medicine’, and dar sustento lit. ‘to give sustenance’), as well as predicates with abstract Ns, close to psychological mental states (e.g., dar ânimo lit. to give encouragement’, dar conforto lit. ‘to give comfort’, dar motivação lit. ‘to give motivation’). In those cases, both eventive and resultative passives can be obtained. Those passives can, therefore, be formed by the verb ser ‘to be’ or the verb ficar ‘to stay’ + the past participle of the V that paraphrases the complex predicate. We will classify these sorts of combinations as instances of transferential event-result complex predicates.

    1. (20)
    1. (a)
    1. ele foi/ficou alimentado, medicado, sustentado
    2. ‘he was/got fed, medicated, sustained’
    1.  
    1. (b)
    1. ele foi/ficou animado/confortado/motivado
    2. ‘he was/was encouraged, comforted, motivated’

Finally, we have shown that light verb dar is even able to combine with Ns that encode physical (e.g., dar dores lit. ‘to give pain’, dar fraqueza lit. ‘to give weakness’, dar vómitos lit. ‘to give vomits’), and psychological sensations (e.g., dar medo lit. ‘to give fear’, dar tristeza lit. ‘to give sadness’). In these cases, the meaning of the complex predicate typically cannot be captured by a corresponding main V; and sometimes there is no corresponding verb at all. These are, therefore, instances of a subtype of dar, which have lost their transference meaning and acquired a causative reading.11 With these predicates, only a resultative passive or a predicate construction with the preposition com ‘with’ + DP arise, as shown in the contrast between (21) and (22). We consider these to be causative complex predicates.

    1. (21)
    1. (a)
    1. *ele foi fraco, triste, bêbedo
    2.   Int. ‘he was weak, sad, drunk’
    1.  
    1. (b)
    1. *ele foi com dores, com fome, com medo
    2.   Int. ‘he/she was in pain, hungry, cold’
    1. (22)
    1. (a)
    1. ele ficou fraco, triste, bêbedo
    2. ‘he became weak, sad, drunk’
    1.  
    1. (b)
    1. ele ficou com dores, com fome, com medo
    2. ‘he became in pain, hungry, cold’

The analyzed data thus allow us to distinguish between three instances of complex predicates with the light verb dar, namely those with (i) a transferential event reading; (ii) a transferential event-result reading; and (iii) a causative reading. Table 2 summarizes the findings.

Table 2: Different values of complex predicates with light verb dar ‘to give’.

transferential event reading transferential event-result reading causative reading
Derived Ns Ns related to Vs √ √ √
Ns related to Ns √
Ns related to ADJs √ √
Underived Ns √

It follows that transferential event reading predicates occur independently of whether the N is related to V, N or ADJ, although transferential event-result predicates occur exclusively with Ns morphologically related to V. A double semantic reading arises from the properties of the deverbal N (Grimshaw 1990). In turn, causative complex predicates are exhibited with N related to Vs and ADJs, as well as with underived Ns. In the next section, we will discuss the semantic features of the three abovementioned complex predicates and account for the variation found in AVPs.

3.2. An account of the variation in AVPs through semantic features

In seminal work by Grimshaw and Mester (1988), light verbs were analyzed as functional elements hosting tense and agreement features but without semantic contribution. However, it has been assumed that, like main verbs, light verbs do contribute to the argument structure of the complex predicate (e.g., Alsina, 1996; Butt, 1995, 2010; Butt & Geuder, 2001; Mohanan, 1994; Rosen, 1989). For EP, this was the approach followed by Gonçalves et al. (2010) concerning complex predicates with light verbs + Ns morphologically related to a V, which denote events.

According to Gonçalves et al. (2010), the formation of a complex predicate depends on feature compatibility between the aspectual properties of the N and the light verb. To form a predication, they propose that light verbs differ from main verbs because they are underspecified for certain aspectual features; as a result, the formation of the complex predicate is essentially an operation of feature compatibility between the light verb and the N morphologically related to a V. Following the aspectual classification from Moens (1987) and Moens and Steedman (1988), Gonçalves et al. (2010), argue that the light verb dar combines with deverbal nouns derived from predicates denoting (i) activities/process (dar um passeio lit. ‘to give a walk’, (ii) accomplishments/culminated processes (dar uma pintura lit. ‘to give a paint’), and (iii) punctuals (dar um espirro lit. ‘to give a sneeze’), but not with (iv) states (*dar um gosto lit. ‘to give a taste’) or (v) culminations (*dar um nascimento lit. ‘to give a birth’). They explain the ungrammaticality of the combination of dar + Ns derived from predicates denoting culminations, by proposing that this light verb is specified with the feature [-change].12

Hence, the data from AVPs we will discuss below pose some challenges to this proposal. It was shown in section 3.1. that light verb dar does not exclusively combine with a deverbal N or a N morphologically related to a V. Therefore, Ns morphologically related to Ns and ADJs (and even some underived Ns) should not be excluded from complex predicate formation.13 On the other hand, it is not adequate to argue that all these Ns preserve the aspectual features of the main verb they are derived from, because they are not always deverbal nominalizations. An alternative hypothesis would be to consider that the N occurring in a complex predicate shares the aspectual features of the verb that can paraphrase the complex predicate, irrespective of their morphological derivation, that is, an approach that accounts for complex predicate formation beyond the morphological nature of the N. In this case, there would not properly be inheritance of aspectual features; they would only share aspectual features. However, as we have seen, the AVP data also exhibits occurrences of predicates that do not have a corresponding main verb. Moreover, we are aware that not all languages allow for a single verb corresponding to a light verb + N combination (see Hale & Keyser 2002, p. 117), and that not all combinations can be paraphrased by a V in the same language.14 To explain these specific cases, next we will discuss an alternative way of understanding the formation of complex predicates.

Suppose that the crucial property to allow the formation of a complex predicate arises from the predicative nature of the N: the N must enter the numeration with a specified [predicative] feature. The aspectual features of the N are then independent of whether it is morphologically related to V, N or ADJ. This approach allows the light verb dar to combine with an underived N, such as Ns denoting psychological and physical sensations, if it is specified for a predicate feature.

In fact, note that the syntactic passivization test used to distinguish between types of complex predicates in the previous section also showed that not all Ns morphologically related to Vs are predicative, i.e., able to form a complex predicate with the light verb dar’. This is the case of combinations such as dar contributo/contribuição lit. ‘to give contribution’, dar colaboração lit. ‘to give collaboration’, dar exemplo lit. ‘to give example’, dar explicação lit. ‘to give explanation’, dar pergunta lit. ‘to give question’, dar resposta lit. ‘to give answer’, among several others that occur in the corpus (see (23)). In these cases, passivization would yield ungrammatical examples in EP (see (24)). Therefore, although these Ns are morphologically related to Vs, they do not share the predicative feature, which is a crucial property to allow the formation of a complex predicate in our analysis.

    1. (23)
    1. (a)
    1. eu
    2. I
    1. vou-lhe
    2. go-3SG.DAT
    1. dar
    2. give
    1. um
    2. an
    1. exemplo
    2. example
    1. concreto (AP)
    2. concrete
    1. ‘I will give you a concrete example.’
    1.  
    1. (b)
    1. ficaram
    2. were
    1. de
    2. of
    1. dar
    2. give
    1. uma
    2. an
    1. resposta (MOP)
    2. answer
    1. ‘They promised to give an answer.’
    1.  
    1. (c)
    1. dar-lhe
    2. give-3SG.DAT
    1. a
    2. the
    1. sua
    2. your
    1. colaboração (STP)
    2. collaboration
    1. ‘…give him/her your collaboration.’
    1. (24)
    1. *ele foi exemplificado/respondido/colaborado
    2. he was exemplified / answered / collaborated

In these cases, the Theme, although an abstract N, is an internal argument of the main V dar; the structure is not one of a complex predicate, but a typical ditransitive structure, as the one with a concrete N.

As already mentioned, according to Gonçalves et al.’s (2010) approach, light verb dar is specified for certain aspectual features. Being specified with the semantic feature [–change], it cannot combine with culminations, which imply a consequent state (*dar um nascimento ‘to give a birth’), but it can combine with culminated processes. Furthermore, dar is specified for the feature [+dynamic], which rules out combinations with states, specified as [–dynamic] (*dar uma estada lit. ‘to give a stay’). The data presented below show that, in AVPs, the light verb dar can combine with culminations, such as dar parto lit. ‘to give birth’ (see (25). Moreover, the combination with states is available in both AVPs and standard EP, as in dar bom cheiro lit. ‘to give good smell’, dar bom gosto lit. ‘to give good taste’, and dar consistência lit. ‘to give consistency’ (see (26) below).

    1. (25)
    1. o que
    2. what
    1. podiam
    2. could
    1. ter
    2. have
    1. feito
    2. done
    1. para
    2. to
    1. evitar
    2. prevent
    1. este
    2. this
    1. acontecimento
    2. event
    1. de a
    2. of the
    1. senhora
    2. lady
    1. dar
    2. give
    1. parto
    2. childbirth
    1. numas
    2. in some
    1. condições
    2. conditions
    1. precárias (MOP)
    2. precarious
    1. ‘What could they have done to prevent the lady from giving birth in precarious conditions?”
    1. (26)
    1. (a)
    1. flor
    2. flower
    1. desse
    2. of.this
    1. mixkitu (…),
    2. mixkitu,
    1. pode-se
    2. can-SE
    1. pisar
    2. crush
    1. também,
    2. also
    1. meter
    2. put
    1. no
    2. in.the
    1. calulu,
    2. calulu
    1. dá
    2. give
    1. bom
    2. good
    1. cheiro
    2. smell
    1. e
    2. and
    1. bom
    2. good
    1. gosto (STP)
    2. taste
    1. ‘The flower of the mosquito-leaf, you can also crush it, put it in the calulu, gives a good smell and a good taste.’
    1.  
    1. (b)
    1. [o milho]
    2. the maize
    1. pisa-se
    2. crushes-SE
    1. junto
    2. together
    1. do
    2. of.the
    1. farelo (…)
    2. bran (…)
    1. para
    2. to
    1. dar
    2. give
    1. consistência
    2. consistency
    1. ao
    2. to.the
    1. farelo (AP)
    2. bran
    1. ‘The maize is crushed together with the bran to give the bran consistency.’

First, let’s consider example (25), which suggests that dar can be combined with a culmination. The event can be interpreted as having no duration; note that it can occur with a temporal expression of punctuality (see (27a)). However, in this case, the predicate could be interpreted as a culmination that has been added to a process, giving rise to a culminated process (see (27b)).15 As a result, (25) can be interpreted as a culminated process, confirming the predictions of Gonçalves et al. (2010).

    1. (27)
    1. (a)
    1. A
    2. the
    1. senhora
    2. lady
    1. deu
    2. gave
    1. parto
    2. childbirth
    1. às
    2. at
    1. 17h.
    2. 17h
    1. ‘The lady gave birth at 17h.’
    1.  
    1. (b)
    1. A
    2. the
    1. senhora
    2. lady
    1. deu
    2. gave
    1. parto
    2. birth
    1. em
    2. in
    1. 5 minutos.
    2. 5 minutes
    1. ‘The lady gave birth in 5 minutes.’

Irrespective of (25) expressing a culmination or a culminated process, it shows that combinations with the light verb dar in AVPs are looser than in the EP grammar, independently of the aspectual class of the V which the N is related to.16 Further examples are provided in (28). In our approach, the predicative nature of the N that combines with the light verb dar in these varieties accounts for more flexible combinations.

    1. (28)
    1. (a)
    1. futuramente
    2. in the future
    1. dar-lhe
    2. give-3SG.DAT
    1. casamento (AP)
    2. marriage
    1. ‘In the future marry her.’
    1.  
    1. (b)
    1. foi um
    2. was an
    1. alarme
    2. alarm
    1. muito
    2. very
    1. grande
    2. big
    1. em
    2. in
    1. São Tomé
    2. São Tomé
    1. e
    2. and
    1. a partir
    2. from
    1. dali
    2. there
    1. eles
    2. they
    1. deram
    2. gave
    1. o
    2. the
    1. eco
    2. echo
    1. suficiente (STP) (=13b)
    2. enough
    1. ‘It sounded a big alarm in São Tomé and from that moment on they paid enough attention to it.’
    1.  
    1. (c)
    1. isso
    2. that
    1. dá
    2. give
    1. entrave
    2. obstacle
    1. a
    2. to
    1. outras
    2. other
    1. coisas (STP)
    2. things
    1. ‘That gets in the way of other things.’
    1.  
    1. (d)
    1. essa
    2. that
    1. possibilidade
    2. possibility
    1. que
    2. that
    1. o
    2. the
    1. reitor
    2. dean
    1. deu
    2. gave
    1. de
    2. of
    1. dar
    2. give
    1. apuramento
    2. clearance
    1. especial
    2. special
    1. para
    2. to
    1. os
    2. the
    1. funcionários. (MOP)
    2. employees
    1. ‘That possibility that the dean has given of granting a special qualification to the employees.’

Let us now turn to the examples in (26). If the combination of the light verb dar with states would not be available in Portuguese, complex predicates of the type shown above would be ruled out. At this point, we should highlight that the prediction of Gonçalves et al.’s (2010) analysis is only confirmed by Ns morphologically related to Vs that denote individual-level predicates. It cannot be extended to the Ns denoting stage-level predicates in (26).

To accommodate the discussed data under a semantic feature account, we propose that the light verb dar is, indeed, unspecified for the feature [change]. Note that, in Gonçalves et al. (2010) approach, the light verb fazer ‘to do’ is also unspecified for the features [change], which would explain the possibility of occurring with culminations. The fact that these two light verbs share semantic features finds empirical support in examples in the AVP where dar is used in combinations that in EP are only allowed with the verb fazer ‘to do’. See the examples in (29), equivalent to fazer uma pergunta lit. ‘to make a question’, ‘to ask’; fazer um resumo lit. ‘to make a summary, ‘to summarize’; and fazer uma reciclagem lit. ‘to make a recycling’, ‘to recycle’.17

    1. (29)
    1. (a)
    1. começo
    2. start
    1. a
    2. to
    1. lhe
    2. 3SG.DAT
    1. dar
    2. give
    1. pergunta (AP)
    2. questions.’
    1. ‘I start to ask him/her questions.’
    1.  
    1. (b)
    1. ele
    2. he
    1. tem
    2. has
    1. que
    2. that
    1. dar
    2. give
    1. um
    2. a
    1. resumo (MOP)
    2. summary
    1. ‘He has to give a summary.’
    1.  
    1. (c)
    1. ano
    2. following
    1. seguinte
    2. year
    1. veio
    2. came
    1. uns
    2. some
    1. estrangeiros
    2. foreigners
    1. da
    2. from
    1. FIFA (…)
    2. FIFA
    1. dar
    2. give
    1. uma
    2. a
    1. reciclagem (STP)
    2. refresher course
    1. ‘The following year, some foreigners from FIFA came to give a refresher course.’

Moreover, we propose that the light verb dar is also unspecified for the feature [dynamic], allowing the combination with episodic states, which, as we have seen, occur exclusively in causative complex predicates (see examples (10), (17), (18)). This possibility only occurs, however, when dar is not a transferential verb; note that in predicates like *dar uma estada lit. ‘to give a stay’, there is no causative reading.

By extending the analysis proposed in Gonçalves et al. (2010) to Ns that are not exclusively related to Vs, we can account for the variation in AVP, considering (i) the predicative feature of the N and (ii) the unspecified features [change] and [dynamic] of the light verb dar. In addition, our approach explains the availability of combinations in EP that are apparently ruled out in an aspectual analysis in which the N preserves the aspectual features of the main verb it is derived from.

The array of Ns that can be specified for the predicative feature as well as the light verb they can combine with varies cross-linguistically. As pointed out by Alba-Sales (2002, p. 51), “for example, in Spanish paseo ‘walk’ combines with dar ‘[to] give’, whereas its Italian counterpart passeggiata appears with fare ‘[to] do’ (cf. English walk, which occurs with take).” Moreover, the author states that, “even within the same language we cannot predict the choice of light verb based on the semantics of the V N[ominalization] alone”, because “diachronically, a given V N[ominalization] may be compatible with certain light verbs at some point in its historical development, but not in others” (see also Acedo-Matellán & Pineda, 2019). Our approach considers the possibility of the N being predicative in a language, and not in others, explaining the variation observed cross-linguistically.

With respect to the semantic features, we suggest that complex predicates with event and event-result readings preserve and share the transferential feature (henceforth [transf]) of the main V dar, as well as the [dynamic] feature. However, the N with an event or a double event-result reading does not correspond to a concrete entity; instead, the N of these complex predicates has a predicate feature, independently of concreteness/abstractness. Also, our proposal distinguishes between complex predicates with transferential event and transferential event-result readings according to the feature [change], which only occurs in the former, where it allows an interpretation of a change of state predicate. Finally, when dar occurs with underived Ns referring to psychological and physical sensations, it completely loses its transference meaning, giving rise to a causative (typically) state interpretation. These predicates share the semantic feature [change] with transferential event-result predicates, but are not dynamic.

Below we summarize the described semantic features of complex predicates with dar.

    1. (30)
    1. (a)
    1. transferential event reading: [+transf, –causative, –change, +dynamic]
    1.  
    1. (b)
    1. transferential event-result reading: [+transf, –causative, +change, +dynamic]
    1.  
    1. (c)
    1. causative reading: [–transf, +causative, +change, –dynamic]

Although we are using AVP data, recall that they do not differ from the data found in standard EP as well as in other languages, irrespectively of the morphological nature of the N. Moreover, the classification of complex predicates based on AVP data is cross-linguistically supported.

Since these different types of predicates are characterized by syntactic properties that have not been explored so far, this will be addressed in the next section.

4. A unified syntactic structure of transferential and causative complex predicates

4.1. Against the unaccusative hypothesis of causative complex predicates

Complex predicates with light verb dar + (underived) Ns expressing psychological and physical sensations have a causative reading, lack a corresponding main verb, and are [–dynamic]. Therefore, Cuervo (2008) argues that these predicates are unaccusative structures.18 One of Cuervo’s arguments in favor of the unaccusative hypothesis of causative predicates with dar resides in the following contrast with the Spanish accusative clitic lo.

    1. (31)
    1. (a)
    1. Le
    2. 3SG.DAT
    1. dieron
    2. gave
    1. apoyoi
    2. support
    1. al
    2. to.the
    1. tecnico?
    2. coach
    1. Si,
    2. Yes,
    1. se
    2. SE
    1. loi
    2. 3SG.ACC
    1. dieron.
    2. gave
    1. ‘Did they give support to the coach? ‘Yes, they gave it to him.’
    1.  
    1. (b)
    1. Le
    2. 3SG.DAT
    1. dimos
    2. gave
    1. miedoi
    2. fear
    1. al
    2. to.the
    1. tecnico?
    2. coach
    1. *Si,
    2. Yes,
    1. se
    2. SE
    1. loi
    2. 3SG.ACC
    1. dimos
    2. gave
    1. ‘Did we frighten the coach? ‘Yes, we did.’
    2. (adapted from Cuervo, 2008, p. 144)

These sequences are also ungrammatical in Portuguese.

    1. (32)
    1. A
    2. the
    1. tempestade
    2. storm
    1. dá
    2. gives
    1. medo
    2. fear
    1. ao
    2. to.the
    1. menino?
    2. boy
    1. *Sim, dá-o.
    2. Yes, give 3SG.ACC
    1. ‘Does the storm frighten the boy? ‘Yes, it does.’

However, although accusative case is arguably not assigned, it is inadequate to consider complex predicates with causative dar unaccusative structures. Unaccusative verbs (like chegar, ‘to arrive’ and ocorrer ‘to occur’) have two important and related properties: they lack an external argument and do not assign accusative case to their internal argument (Burzio, 1986). An argument against the unaccusative hypothesis of causative structures with the light verb dar is the fact that there seems to be an argument Cause/Stimulus, which can be replaced by the demonstrative isso ‘this’, when this argument bears a [–ANIM] feature (cf. (33–34)), and by a personal pronoun, when it is [+ANIM] (cf. (35)).

    1. (33)
    1. (a)
    1. paludismo
    2. malaria
    1. é
    2. is
    1. a
    2. a
    1. febre,
    2. fever,
    1. dá
    2. gives
    1. fraqueza (STP) (= (17c))
    2. weakness
    1. ‘Malaria is a fever, it makes you weak.’’
    1.  
    1. (b)
    1. isso
    2. that
    1. dá
    2. gives
    1. fraqueza
    2. weakness
    1. ‘That gives weakness.’
    1. (34)
    1. (a)
    1. dava
    2. gave
    1. medo
    2. fear
    1. pessoa
    2. person
    1. sair
    2. leave
    1. para
    2. to
    1. ir
    2. go
    1. na
    2. in.the
    1. discoteca (AP) (= (13b))
    2. disco
    1. ‘One would be(come) afraid to go to the disco.’
    1.  
    1. (b)
    1. isso
    2. that
    1. dava
    2. gave
    1. medo
    2. fear
    1. ‘That was frightening.’
    1. (35)
    1. (a)
    1. essa
    2. that
    1. do
    2. from
    1. Chibuto
    2. Chibuto
    1. está
    2. is
    1. me
    2. 1SG.DAT
    1. dar
    2. give
    1. um pouco de
    2. some
    1. esforço (MOP)
    2. effort
    1. ‘That one from Chibuto is giving me a bit of a hard time.’
    1.  
    1. (b)
    1. ela
    2. she
    1. está
    2. is
    1. me
    2. 1SG.DAT
    1. dar
    2. give
    1. um pouco de
    2. some
    1. esforço
    2. effort
    1. ‘She is giving me a bit of a hard time.’

The data discussed above lead to two important conclusions: on the one hand, bare nouns describing psychological or physical sensations in causative predicates with dar are Theme DOs and not Theme Subjects; on the other hand, the light verb is not an unaccusative.

But if causative complex predicates with dar are not unaccusative structures, how can we explain the ungrammaticality of accusative cliticization presented in Spanish? The contrast of grammaticality observed in (31) may arise from the different argument status of the N that combines with dar: while in (31a), the N apoyo seems to be a true argument of the V, the N miedo, being probably a non-argument, is reanalyzed as part of the predicate.19 The same may apply to Portuguese. Therefore, we propose that, due to this different argumenthood of the Theme, this sort of combinations of dar exhibits some degree of lexicalization. It is well known that lexicalization is a broad notion that may include light verbs. Bacelar do Nascimento (2013, pp. 239–240) considers that Portuguese dar occurs in “semi-transparent expressions”, such as dar razão lit. ‘to give reason’ or dar valor, lit. ‘to give value’.

Although these expressions maintain part of the lexical meaning of their elements, they exhibit syntactic properties that allow us to distinguish them from fully lexicalized expressions. One of these properties is the possibility of causative complex predicates with dar to occur not only with bare nouns, but also with definite and indefinite articles, as shown in (17), repeated for convenience in (36), and with quantifiers, as shown in (37).

    1. (36)
    1. (a)
    1. dizem
    2. say
    1. que
    2. that
    1. isso
    2. that
    1. fermentar
    2. ferment
    1. é
    2. is
    1. que
    2. that
    1. dá
    2. gives
    1. a
    2. the
    1. bebedeira (AP)
    2. drunkenness
    1. ‘They say that fermentation is what makes people drunk.’
    1.  
    1. (b)
    1. você
    2. you
    1. vê
    2. see
    1. só
    2. just
    1. ele
    2. he
    1. a
    2. to
    1. chorar
    2. cry
    1. dá
    2. gives
    1. uma
    2. a
    1. tristeza (AP)
    2. sadness
    1. ‘You just see him crying, it makes you sad.’
    1.  
    1. (c)
    1. paludismo
    2. malaria
    1. é
    2. is
    1. a
    2. a
    1. febre,
    2. fever,
    1. dá
    2. gives
    1. fraqueza (STP)
    2. weakness
    1. ‘Malaria is a fever, it makes you weak.’
    1. (37)
    1. (a)
    1. é
    2. is
    1. uma
    2. a
    1. coisa
    2. thing
    1. que
    2. that
    1. dá
    2. gives
    1. muito
    2. much
    1. prazer
    2. pleasure
    1. em
    2. in
    1. fazer
    2. do
    1. isso (STP)
    2. that
    1. ‘It’s something that’s a real pleasure to do.’
    1.  
    1. (b)
    1. uma
    2. a
    1. [filha]
    2. daughter
    1. em
    2. in
    1. casa
    2. home
    1. que
    2. who
    1. me
    2. 1SG.DAT
    1. dá
    2. gives
    1. muito
    2. much
    1. trabalho (AP)
    2. work
    1. ‘I have a [daughter] at home who gives me a lot of work.’
    1.  
    1. (c)
    1. no
    2. in.the
    1. muro
    2. wall
    1. também (…)
    2. also
    1. dá
    2. gives
    1. muita
    2. a lot of
    1. frialdade (STP)
    2. coldness
    1. ‘The wall also gives a lot of coldness.’

Moreover, we can modify structures such as dar raiva lit. ‘to give anger’ and dar febre lit. ‘to give fever’, with both psychological Ns and Ns denoting physical sensations, to structures with adjectives: dar uma grande raiva lit. ‘to give a great anger’ and dar uma febre alta lit. ‘to give a high fever’ (see also Bruening (2020), for English). The possibility of modifying the Theme is an essential piece of evidence in favor of the semi-lexicalization of these combinations. Note further that, in the same context, other Ns can also occur (dar raiva / fúria / ira lit. ‘to give anger’), which also supports the claim that these combinations are not fully lexicalized.

To sum up, in causative complex predicates with dar, the argumental status of the DO is arguable. As a result, although these semi-lexical combinations of V + underived Ns does allow some manipulations, as shown above, they do not allow the cliticization of the DO.

4.2. The argumental status of the IO and their thematic role

The IOs in causative complex predicates with dar + Ns of psychological and physical sensations have been characterized in the literature as an Experiencer related to stative predicates.20 Cuervo (2008), for instance, uses this specific meaning to propose that, in Spanish, when Experiencers have the form of datives, they are very often characterized by their initial position, doubled by a clitic, and they do not admit a post-verbal position (although there is interspeaker variation).

    1. (38)
    1. (a)
    1.   Al
    2.   to.the
    1. técnico
    2. coach
    1. le
    2. 3SG.DAT
    1. dan
    2. give
    1. rabia
    2. fury
    1. las
    2. the
    1. protestas.
    2. complaints.nom
    1.   ‘The complaints make the coach furious.’
    1.  
    1. (b)
    1. *Las
    2.   the
    1. protestas
    2. complaints.nom
    1. dan
    2. give
    1. miedo
    2. fear
    1. al
    2. to.the
    1. tecnico.
    2. coach
    1.   ‘The coach is afraid of the complaints.’
    2. (Cuervo, 2008, p. 140)

Based on these properties, Cuervo (2008, p.146) argues that in (38a) the dative is merged outside the VP, i.e., it is not an internal argument, and it is a high applied dative: “the dative experiencer is licensed as the specifier of a High Applicative head, that is, an applicative that takes the vP as its complement, not a DP (Pylkkanen, 2008).” Moreover, she states that VoiceP is not projected; therefore, there is no accusative case or (agentive) external argument. In her approach, predicates with dar + psychological Ns are unaccusative structures, and dative experiencers exhibit subject properties.21

Jiménez-Fernández and Rozwadowska (2017), on the other hand, propose that, in Spanish, the distribution of dative Experiencers in the sentence is influenced by both argument structure and information structure. According to them, the order OVS is preferable only when it is used as a reply to an open question, such as What’s up? What is happening? Therefore, (39b) is not an appropriate answer to these questions (see also Fábregas, Jiménez-Fernández & Tubino 2017).

    1. (39)
    1. (a)
    1.   A
    2.   to
    1. Luis
    2. Luis
    1. le
    2. 3SG.DAT
    1. interesa
    2. interest
    1. la
    2. the
    1. sintaxis.
    2. syntax
    1.  
    1. (b)
    1. #La sintaxis le interesa a Luis.
    2.   ‘Luis is interested in syntax.’
    3. (Jiménez-Fernández & Rozwadowska, 2017, p. 3)

Therefore, they claim that dative experiencers are preverbal not because they are subjects, but because they move to Spec, TP. From the point of view of argument structure, the dative is an internal argument, “generated in VP, in a position higher than the nominative subject, and hence it is the closest candidate to move to spec-TP and satisfy the EPP under T” (Jiménez-Fernández & Rozwadowska, 2017, p. 236). On the other hand, from the perspective of information-structure, dative Experiencers “occur first only in two situations, namely when they are part of the broad focus that the whole sentence performs (…) or when they function as topic.” (Ibidem).

Portuguese differs from Spanish in this respect; the dative in the initial position cannot even be understood as a subject. This position is mainly accepted for contrastive topics, for instance in answers to parallel questions, where the focus is the object.

    1. (40)
    1. - O que interessa ao João? O que interessa à Ana?
    2. ‘What does interest John? What does interest Anne?’
    3. - Ao João interessa a fonologia, à Ana interessa a sintaxe.
    4. ‘To John, phonology interests, to Anne syntax interests.’

Thus, this initial position in Portuguese is not a reason to consider an Experiencer dative as a subject nor as a non-argument dative in causative complex predicates.22

Let us return to the status of IO with light verbs in general. Analyzing the light verb dar in EP transferential predicates of the type illustrated in (2c), Brito (2009) argues that it loses the capacity of selecting an argument dative, due to semantic bleaching. According to the author in dar uma pintura às paredes lit. ‘to give a painting to walls’ (‘to paint the walls’), the DP uma pintura ‘a painting’ is not used as a concrete object of possession transference between two individuals, but instead as a (simple) event noun and therefore refers to something more abstract than generally occurs with the full verb dar. On the other hand, the DP paredes ‘walls’ is not a human entity able to be the Recipient of transfer of possession. Therefore, Brito considers that in structures such as (2c) the IO is a Beneficiary (see Choupina & Brito 2018, for a similar proposal). Wittenberg et al. (2014) and Wittenberg (2016), on the other hand, propose that similar examples in English are instances of transfer of metaphorical possession, whereas Butt & Geuder (2001, p. 342) argue that, in these cases, there is only an “application of an action to an entity” (see section 1). None of these authors treat the IO of light verbs as a non-argument dative. And, in fact, in examples like dar uma pintura às paredes lit. ‘to give a painting to walls’ (‘to paint the walls’), the DP as paredes ‘the walls’ are both the Recipient of the light verb dar and the affected Theme (Patient) of the main verb which the N is morphologically related to.23 Therefore, one can argue that there is a mixed thematic role assignment: the IO is not a Beneficiary (nor a non-argument IO), as proposed by Brito (2009) and Choupina and Brito (2018), but instead an internal argument with the thematic role of Recipient-Patient.24

Moreover, the argument / non-argument nature of the IO can be empirically tested. Gutiérrez Ordóñez (1999, p. 1886), argues that, in Spanish, only participles of verbs that subcategorize argument datives can be combined with an IO (a-DP). Gonçalves (2016) adapts this test to the Portuguese data, against Brito (2009), showing that the participial structures are grammatically identical to the main verb dar (41a), irrespective of the light / full nature of the verb and the animacy of the IO.

    1. (41)
    1. (a)
    1. O
    2. the
    1. livro
    2. book
    1. dado
    2. given
    1. ao
    2. to.the
    1. amigo
    2. friend
    1. é
    2. is
    1. uma
    2. a
    1. relíquia.
    2. relic
    1. ‘The book given to his friend is a relic.’
    1.  
    1. (b)
    1. O
    2. the
    1. conselho/ânimo
    2. advice
    1. dado
    2. given
    1. ao
    2. to.the
    1. amigo
    2. friend
    1. ajudou-o
    2. helped-3SG.ACC
    1. muito.
    2. much
    1. ‘The advice given to his friend helped him a lot.’
    1.  
    1. (c)
    1. O
    2. the
    1. abraço/beijo
    2. hug/kiss
    1. dado
    2. given
    1. à
    2. to.the
    1. rapariga
    2. girl
    1. surpreendeu-a.
    2. surprised-3SG.ACC
    1. ‘The hug/kiss given to the girl surprised her.’
    1.  
    1. (d)
    1. A
    2. the
    1. pintura
    2. painting
    1. dada
    2. given
    1. às
    2. to.the
    1. paredes
    2. walls
    1. lascou.
    2. chipped
    1. ‘The painting given to the walls chipped.’

However, the past participle test (Gutiérrez Ordóñez, 1999) used with causative complex predicates shows that IOs behave differently from the IOs of transferential complex predicates. As shown in (42), participial structures are ruled out.

    1. (42)
    1. (a)
    1. *A
    2.   the
    1. dor/fraqueza
    2. pain/weakness
    1. dada
    2. given
    1. à
    2. to.the
    1. Maria
    2. Mary
    1. impediu-a
    2. prevented-3SG.ACC
    1. de
    2. from
    1. sair
    2. leave
    1. de
    2. the
    1. casa.
    2. house
    1.   Int. ‘The pain / weakness given to Mary prevented her from leaving the house.’
    1.  
    1. (b)
    1. *A
    2.   the
    1. tristeza/preguiça
    2. sadness/laziness
    1. dada
    2. given
    1. à
    2. to.the
    1. Maria
    2. Mary
    1. impediu-a
    2. prevented-3SG.ACC
    1. de
    2. from
    1. sair
    2. leave
    1. de
    2. the
    1. casa.
    2. house
    1.   Int. ‘The sadness / laziness given to Mary prevented her from leaving the house.’

On the one hand, these examples show that the syntactic relation of the dative with a transferential ditransitive verb is similar no matter whether this verb is a full ditransitive or a light verb. On the other hand, the relation appears to be different when the verb dar is specified with the causative feature.

Note also that only IOs of transferential event and double event-result readings allow the possibility of both DO and IO to be clefted as a constituent, differently from a ditransitive structure with dar as a main verb (see (43c–44c)). Gonçalves et al. (2010) argue that this possibility points to the fact that the IO is associated with the properties of categorial selection of the deverbal N.

    1. (43)
    1. (a)
    1. O
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. um
    2. a
    1. livro
    2. book
    1. aos
    2. to.the
    1. alunos.
    2. students
    1. ‘The teacher gave a book to the students.’
    1.  
    1. (b)
    1. O
    2. the
    1. que
    2. that
    1. o
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. aos
    2. to.the
    1. alunos
    2. students
    1. foi
    2. was
    1. um
    2. a
    1. livro.
    2. book
    1. ‘What the teacher gave to the students was a book.’
    1.  
    1. (c)
    1. ??O
    2.     the
    1. que
    2. that
    1. o
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. foi
    2. was
    1. um
    2. a
    1. livro
    2. book
    1. aos
    2. to.the
    1. alunos.
    2. students
    1. ‘What the teacher gave was a book to the students.’
    1. (44)
    1. (a)
    1. O
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. um
    2. a
    1. grande
    2. big
    1. ralhete /
    2. scold /
    1. uma
    2. a
    1. preciosa
    2. precious
    1. ajuda
    2. help
    1. aos
    2. to.the
    1. alunos.
    2. students
    1. ‘The teacher gave a big scold / a precious help to the students.’
    1.  
    1. (b)
    1. O
    2. the
    1. que
    2. that
    1. o
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. aos
    2. to.the
    1. alunos
    2. students
    1. foi
    2. was
    1. um
    2. a
    1. grande
    2. big
    1. ralhete/
    2. scold/
    1. uma
    2. a
    1. preciosa
    2. precious
    1. ajuda.
    2. help
    1. ‘What the teacher gave to the students was a big scold / a precious help.’
    1.  
    1. (c)
    1. O
    2. the
    1. que
    2. that
    1. o
    2. the
    1. professor
    2. teacher
    1. deu
    2. gave
    1. foi
    2. was
    1. um
    2. a
    1. grande
    2. big
    1. ralhete/
    2. scold/
    1. uma
    2. a
    1. preciosa
    2. precious
    1. ajuda
    2. help
    1. aos
    2. to.the
    1. alunos.
    2. students
    1. ‘What the teacher gave was a big scold / a precious help to the students.’

If one considers causative complex predicates, which, as we have shown, mainly occur with underived Ns of psychological and physical sensations, the possibility of simultaneous extraction is again ruled out (see (45c)).

    1. (45)
    1. (a)
    1. A
    2. the
    1. infeção
    2. infection
    1. respiratória
    2. respiratory
    1. deu
    2. gave
    1. fraqueza/tristeza
    2. weakness/sadness
    1. à
    2. to.the
    1. Maria.
    2. Maria
    1. ‘The respiratory infection made Maria weak/sad.’
    1.  
    1. (b)
    1. O
    2. the
    1. que
    2. that
    1. a
    2. the
    1. infeção
    2. infection
    1. respiratória
    2. respiratory
    1. deu
    2. gave
    1. à
    2. to.the
    1. Maria foi fraqueza/ tristeza.
    2. Maria was weakness/sadness
    1. ‘What the respiratory infection did to Maria was make her weak/sad.’
    1.  
    1. (c)
    1. */??O
    2.        the
    1. que
    2. that
    1. a
    2. the
    1. infeção
    2. infection
    1. respiratória
    2. respiratory
    1. deu
    2. gave
    1. foi
    2. was
    1. fraqueza/tristeza
    2. weakness/sadness
    1. à
    2. to.the
    1. Maria
    2. Maria
    1. ‘What the respiratory infection did to Maria was make her weak/sad.’

Therefore, we believe that the source of rejection of participial structures as well as the oddness of simultaneous extraction are the same: in causative complex predicates, there is no possibility of the IO being somehow dependent on the DO. Moreover, the ungrammaticality of (45c) says nothing about the argument nature of the IO; if so, one should also argue that the oddness of (43c) suggests that the IO of a ditransitive full verb is not an internal argument of the V, an idea that we reject altogether. Therefore, we consider that (45c) and accusative cliticization (cf. section 4.1) are rejected for the same reason, namely that causative complex predicates with underived Ns are more lexicalized than transferential complex predications.

An additional piece of evidence in favor of the degree of lexicalization is the fact that AVPs do not exhibit the same amount of variation in the expression of the IO in ditransitive structures with the full verb, on the one hand, and in ditransitive structures with transferential and causative complex predicates, on the other. Although the corpus data exhibit complex predicates in the form of DOCs or DPCs introduced by prepositions other than the Case marker a, these structures are restricted to transferential predicates with an event reading, and they only occur to a very limited extent. 79% of the analyzed data involving complex predicates with full DPs correspond to both transferential and causative structures with the IO introduced by the Case marker a (98/124 occurrences) (see (8a) repeated for convenience in (46a)). The corpus only exhibits 10 DOCs (8%) (see (46b)) and 16 DPCs introduced by the preposition em and para, with 8 occurrences each (13%), exclusively with transferential predicates (see (10b) and (28d), repeated for convenience in (46c–d)).

    1. (46)
    1. (a)
    1. meu
    2. my
    1. sonho
    2. dream
    1. seria
    2. would be
    1. dar
    2. give
    1. educação
    2. education
    1. aos
    2. to.the
    1. meus
    2. my
    1. filhos (MOP)
    2. children
    1. ‘My dream would be to give education to my children’.
    1.  
    1. (b)
    1. não
    2. not
    1. tem
    2. has
    1. estado
    2. been
    1. a
    2. to
    1. dar
    2. give
    1. professor
    2. teacher
    1. muita
    2. a lot of
    1. consideração (STP)
    2. consideration
    1. ‘has not been giving the professor much consideration’.
    1.  
    1. (c)
    1. estão
    2. are
    1. a
    2. to
    1. dar
    2. give
    1. o
    2. the
    1. abate
    2. cut
    1. em
    2. in
    1. muitas
    2. many
    1. árvores. (AP)
    2. trees
    1. ‘they are cutting down many trees.’
    1.  
    1. (d)
    1. essa
    2. that
    1. possibilidade
    2. possibility
    1. que
    2. that
    1. o
    2. the
    1. reitor
    2. dean
    1. deu
    2. gave
    1. de
    2. of
    1. dar
    2. give
    1. apuramento
    2. clearance
    1. especial
    2. special
    1. para
    2. to
    1. os
    2. the
    1. funcionários. (MOP)
    2. employees
    1. ‘that possibility that the dean has given of granting a special qualification to the employees.’

We found no causative complex predicates exhibiting a divergent syntactic structure from EP. Hence, we must conclude that the expression of Experiencers is not subject to variation in AVPs, contrary to Recipients of (full) core dative verbs (e.g., Gonçalves, Duarte & Hagemeijer 2022). Moreover, variation is almost inexistent with Recipients of transferential complex predicates with double event-result reading, being observed exclusively with pronominal NPs, particularly in STP (see (9a) and (12a), repeated in (47a-b)).25

    1. (47)
    1. (a)
    1. Ele
    2. He
    1. está
    2. is
    1. me
    2. 3SG.DAT
    1. dar
    2. give
    1. muito
    2. a lot of
    1. ânimo. (STP)
    2. encouragement
    1. ‘He is strongly encouraging me.’
    1.  
    1. (b)
    1. essa
    2. that
    1. galinha
    2. chicken
    1. que
    2. that
    1. deu
    2. gave
    1. ela
    2. she
    1. encorajamento (STP)
    2. encouragement
    1. ‘That chicken that gave her encouragement…’

In the next section, we will present a unifying syntactic structure of complex predicates with the light verb dar.

4.4. The syntactic structure of transferential and causative complex predicates

Duarte et al. (2009) and Gonçalves et al. (2010) argue that the EP light verb dar has a predicative status (and not a functional one) due to three main properties: (i) it preserves the argument structure of the homophonous main verb dar ‘give’, that is, an external argument and two internal arguments; (ii) it is responsible for the semantic selection of the external argument; and (iii) it preserves part of the meaning of the corresponding main verb. We agree with the authors regarding transferential predicates but we propose that causative dar does not share properties (ii) and (iii).

As for the external argument with the thematic role of Cause / Stimulus, it can be [±ANIM] (see (13a) and (13b), for instance), and may not be a controller, since it generally does not license adverbs of intentionality, as for instance propositadamente ‘purposely’ and intencionalmente ‘intentionally’. This is explained by the feature [+cause] which is valued for the light verb dar when it combines with a predicative nominal expression of psychological and physical sensations.

As for the IO, it is an internal argument of the light verb, but crucially it is an Experiencer with causative complex predicates, not a Recipient-Patient, as in transferential complex predicates. However, we consider that the conditions for Case assignment are the same.26 As the IO of core dative verbs in Portuguese (and Romance languages), the IO of complex predicates in both AVPs and EP is a(n extended) DP (and not a PP). It occurs internally to a K(ase) projection, whose head is the functional preposition (Case marker) a or a dative clitic (Gonçalves, 2016; Gonçalves, Duarte & Hagemeijer, 2022).

This unified syntactic proposal is reminiscent of Arad (1998), who argues that the semantic differences in this source of predicates have no syntactic repercussions. In this sense, our proposal differs from other proposals that invoke supplementary functional categories above the VP, like an Exp(erience) P(hrase) (e.g., Viñas-de-Puig, 2008, 2014), and an Appl(icative) P(hrase) (e.g., Cuervo, 2010; Bruening, 2015, Folli & Harley 2013, among others).

We propose the following syntactic representation for complex predicates with the light verb dar, irrespective of the different semantic features of the predicate.

    1. (48)

The IO is projected in Spec, VP, and does not need to move for Case assignment purposes; its Case is assigned internally to KP27, a functional projection whose head is the Case marker a ‘to’. On the other hand, the Theme DP raises to Spec, vP to valuate its accusative case. The external argument, an Agent or Cause, depending on whether the complex predicate is transferential or causative, is projected as Spec, VoiceP (cf. Alexiadou, 2014, Kratzer, 1996,).28

Finally, we would also like to present a couple of complex predicates that occur in the corpus and are distinct from the type discussed above, since they lack an overtly expressed Cause/Stimulus, such as dar preguiça lit. ‘to give laziness’, dar vontade lit. ‘to give desire’, and dar impressão lit. ‘to give impression’.

    1. (49)
    1. (a)
    1. quando
    2. when
    1. cheguei
    2. arrived
    1. na
    2. in.the
    1. 8.ª classe,
    2. 8th grade,
    1. me
    2. 1SG.DAT
    1. dava
    2. gave
    1. preguiça
    2. laziness
    1. de
    2. to
    1. estudar. (STP)
    2. study
    1. ‘When I got to the 8th grade, I was too lazy to study.’
    1.  
    1. (b)
    1. não
    2. not
    1. me
    2. 1SG.DAT
    1. dá
    2. gives
    1. muita
    2. much
    1. vontade. (MOP)
    2. will
    1. ‘I don’t really feel like it.’
    1.  
    1. (c)
    1. dá
    2. gives
    1. a
    2. the
    1. impressão
    2. impression
    1. que
    2. that
    1. quem
    2. who
    1. não
    2. not
    1. faz… (STP)
    2. does
    1. ‘It gives the impression that those who don’t do it…’

These data only exhibit singular Themes, but examples like (50b), with plural nouns, are also common in EP.

    1. (50)
    1. (a)
    1. Estava
    2. was
    1. a
    2. to
    1. trabalhar
    2. work
    1. e,
    2. and
    1. subitamente,
    2. suddenly
    1. deu-me
    2. gave 1SG.DAT
    1. fome/medo/frio.
    2. hunger/fear/cold
    1. ‘I was working and suddenly it made me hungry / afraid / cold.’
    1.  
    1. (b)
    1. Estava
    2. was
    1. a
    2. to
    1. trabalhar
    2. work
    1. e,
    2. and
    1. subitamente,
    2. suddenly
    1. deu-me
    2. gave-1SG.DAT
    1. tonturas/vómitos/nervos.
    2. dizzy/vomiting/nerves
    1. ‘I was working and suddenly it made me dizzy / puke / nervous.’

If the light verb dar would correspond to an unaccusative structure, as suggested by Cuervo (2008), the lack of agreement in (50b) is unexpected.29 Besides, nominative case cannot be assigned and there is no obligatory agreement (see (51–52)), contrary to what is observed with unaccusative verbs, as in (53). Finally, (51–52) still exhibit Theme direct objects, but no Theme subjects, as would be the case of true unaccusative verbs, as in (53).

    1. (51)
    1. (a)
    1. Deu-me
    2. gave-1SG.DAT
    1. fome /
    2. hunger/
    1. *ela
    2.   she
    1. deu-me
    2. gave-1SG.DAT
    1. fome.
    2. hunger
    1. ‘S/he/it gave me hunger’
    1.  
    1. (b)
    1. Deu-me
    2. gave-1SG.DAT
    1. nervos /
    2. nerves /
    1. *eles
    2.   they
    1. deram-me
    2. gave-1SG.DAT
    1. nervos.
    2. nerves
    1. ‘They made me nervous’.
    1. (52)
    1. (a)
    1. Deu-me
    2. gave-1SG.DAT
    1. vómitos.
    2. vomiting
    1. ‘It made me vomit.’
    1.  
    1. (b)
    1. */??Deram-me
    2. gave-1SG.DAT
    1. vómitos.
    2. vomiting
    1. ‘They made me vomit.’
    1. (53)
    1. (a)
    1. Ocorreu-me
    2. occurred-1SG.DAT
    1. uma
    2. an
    1. ideia /
    2. idea /
    1. ela
    2. it
    1. ocorreu-me.
    2. occurred-1SG.DAT
    1. ‘An idea occurred to me. / It occurred to me.’
    1.  
    1. (b)
    1. Aconteceu-me
    2. happened-1SG.DAT
    1. um
    2. a
    1. percalço /
    2. mishap /
    1. ele
    2. it
    1. aconteceu-me.
    2. happened-1SG.DAT
    1. ‘A mishap happened to me. / It happened to me.’

In this type of complex predicate, the light verb dar does not display an external argument (the Stimulus / the Cause) and gets a strong stative meaning. We leave this open for future research.

5. Final remarks

In this paper, we have studied complex predicates with the light verb dar based on corpus data from AVPs.

Having started from classical views on light verbs, our analysis has revealed that not only deverbal nouns but also denominal and deadjectival nouns and even non-derived nouns may combine with the light verb dar. To explain these combinations, we argue that the N must enter the numeration with a [+predicative] feature. This property varies cross-linguistically, which explains the possibility of some combinations being allowed in a language (or variety) and excluded from others, as well as the existence of some combinations that, although involving an abstract N (and being paraphrased by a main V), cannot be understood as complex predicates.

Also, this broad notion of predicative noun is related to the existence of distinct subclasses of the light verb dar, which we characterize by the following three readings: (i) a transferential event reading; (ii) a transferential event-result reading; and (iii) a causative stative reading. In terms of semantic features, they may be distinguished by the features [transf] and [dynamic], which are common to (i) and (ii); [cause], exclusive to (iii); and [change], shared by (ii) and (iii). We have used passivization as a syntactic test to empirically support this classification.

We further argue that these semantic features and the predicative status of the N explain why some combinations with the light verb dar are allowed in AVPs but not in EP. Thus, following Gonçalves et al. (2010), we argue that complex predicate formation occurs in syntax. However, instead of focusing on the aspectual features of the predicative N, which, as we have shown is not exclusively deverbal, we consider that the observed microvariation can be captured by assuming that the light verb dar is unspecified for the features [change] and [dynamic].

In what concerns the syntactic structure of these complex predicates, we have shown that there is no empirical motivation to argue that causative complex predicates are unaccusative structures. Building on Arad (1998), we propose that the syntactic structure is rather similar in transferential and causative complex predicates, since both are ditransitive structures. Nevertheless, three important differences should be emphasized: (i) the causative light verb dar has the feature [+cause] (and not [+transf], as in transferential predicates); (ii) the causative complex predicate encodes a non-dynamic situation, typically with a stative reading; and (iii) the IO selected by the causative dar is an Experiencer (and not a Recipient-Patient, as in transferential predicates).

Finally, by analyzing the prototypical core dative verb in complex predicates, where it has the status of a light verb, our proposal leads to a better understanding of the expression of the IO in AVPs. We found that with respect to light verb constructions AVPs do not exhibit the same degree of variation, when considering the distinction between Recipients of a main verb, Recipients-Patients, and Experiencers. Complex predicates are, by their very nature, closer to lexicalized expressions, being less prone to modifications. This can be observed in particular with causative complex predicates, which do not easily allow cliticization of the DO or Theme passivization. This is also suggested by the greater convergence of AVPs towards EP in the expression of Recipients-Patients, and by the total absence of divergence in the expression of Experiencers. As has been shown in several recent papers, AVPs clearly tend to converge with EP in some respects, showing that their syntactic properties are highly driven and constrained by Universal Grammar (UG) (see on this matter Gonçalves, Duarte & Hagemeijer, 2022).

As usual, more work on this topic is necessary to fully understand the complexity of complex predicates.

Acknowledgements

The research in this paper was developed within project PALMA (Possession and location: microvariation in African varieties of Portuguese, PTDC/LLT-LIN/29552/2017), funded by the Portuguese Foundation for Science and Technology (FCT). A preliminary version of this paper was presented at the Colloquium Our ID, Lisbon, 19–20 January 2023. We thank the audience for their valuable comments and suggestions. We are also grateful to Tjerk Hagemeijer for the encouragement to address this topic and for the reading of an earlier version of the text, and to the three anonymous reviewers, whose suggestions greatly improved the final version of this paper.

Competing Interests

The authors have no competing interests to declare.

Notes

  1. Bearing this in mind, Choupina and Brito (2018) consider that the monotransitive dar (as in dar um espirro ‘to sneeze’) in EP corresponds to a higher degree of light verb bleaching since the argument structure of the full verb is changed. [^]
  2. Note also that Rothstein (2004, p. 184), among others, claims that light verbs are one of the contexts that contribute to the expression of telicity, i.e., they contribute to the expression of a delimited event or the endpoint of an event. This explains the reason why, in light verb combinations, the NP whose nominal head is derived from verb roots that denote processes refer to a limited portion of that process, i.e., it conveys bounded events, that may be countable and pluralizable (see (i)).
      1. (i)
      1. O João fez duas caminhadas (durante a semana)
      2. ‘John took two walks during the week’
      3. (adapted from Choupina & Brito, 2018)
    [^]
  3. Considering data from Brazilian Portuguese, Rassi and Vale (2013) argue that causative uses of dar are also available with Ns denoting properties or characteristics (see (i) below). However, those combinations are typically deadjectival or morphologically related to an adjective, such as celeridade ‘celerity’, afetividade ‘affectivity’, amplitude ‘range’, autenticidade ‘authenticity’, and still preserve the transference reading, which causative structures do not exhibit, as we will argue in sections 3 and 4.
      1. (i)
      1. O advogado deu celeridade ao processo.
      2. ‘The lawyer expedited the process.’
    [^]
  4. According to Cuervo, following Marantz (1993) and Pylkkänen (2002, 2008), Spanish ditransitive structures with clitic doubling are applicative constructions, i.e., DOCs, and the IO is an applied constituent. [^]
  5. We follow Rio-Torto (2013, p. 63) in the following assumption: “Even if a lexeme has its Latin correspondent, if this lexeme shows morphological constituents and a formation pattern that are active in Portuguese, it is considered a derived one. (…) Considering the patterns that form action nouns from verbs with the suffix -ção, dedicação [from Lat. Dedicatione] is perfectly framed according to the parameters of word formation of Portuguese” (our translation). [^]
  6. Note, however, that this verb is typically used in Portuguese with the meaning of nodding off. [^]
  7. Scher (2004) addressed this sort of light verb combination in Brazilian Portuguese in the framework of Distributed Morphology, arguing that the suffix -ada has a nominal origin, even if this noun has an intermediate verbal form: it is the case of cabeçada, in dar uma cabeçada ‘to give a headbutt’, passada in dar uma passada ‘to take a step’. These nominals have no Asp and the verbal head is empty; they are simple event nominals, countable and pluralizable. [^]
  8. See Alonso-Ramos (1998, p. 403): “(…) tout nom prédicatif n’ est pas nécessairement un nom abstrait.” [^]
  9. Using the terminology of “construction à verbe support (CVS)” and not “light verb”, Alonso-Ramos (1998), for French, arrives at a similar conclusion: “Si l’on se limite à des noms dérivés morphologiquement d’un verbe ou à des noms ayant une structure d’arguments dans le sens de Grimshaw (1990), alors la quantité de noms qui ne pourraient pas entrer dans une CVS serait énorme.” (p. 403). [^]
  10. Our approach differs from that of Folli and Harley (2013). In a paper on Italian complex predicates formed by combining the nominalization -ata + a light verb, the authors argue that the lack of passivization can be used as a diagnosis for light verb and complex predicate identification (see also Folli & Harley, 2007). The authors argue that the grammaticality contrast observed in (i–ii) arises from the fact that only lexical Vs can have a passive participle form.
      1. (i)
      1. *Una torta è stata fatta da Gianni.
      2. a cake has been made by Gianni
      3. ‘A cake made by Gianni.’
      1. (ii)
      1. *Una letta è stata data a Kant da Gaia.
      2. a reading has been given to Kant by Gaia
      3. ‘A reading was given to Kant by Gaia.’ (see Folli & Harley 2013, p. 108)
    However, note that, in (ii), the authors use a nominalization from an incremental Theme V, that is, the verb leggere ‘to read’. Although “incremental theme verbs may sometimes show non completive readings even with quantized objects” (Rappaport-Hovav & Levin 2008, p. 149), as in example (ii), the V give is non-incremental by its very nature. Its co-occurrence with a nominalization from an incremental theme V leads to ungrammatical sentences, not because of the light V status, as suggested by Folli and Harley (2013), but because the implicature of completion holds with respect to the nominalization. [^]
  11. Note also that only transferential complex predicates involving (concrete) Ns morphologically related to Vs and Ns allow a reverse structure with the V receber ‘to receive’ (cf. (i)). That structure is ruled out with causative complex predicates (cf. (ii)).
      1. (i)
      1. Ele
      2. He
      1. recebeu
      2. received
      1. a
      2. the
      1. indemnização / a
      2. compensation / the
      1. medicação / uma
      2. medication / a
      1. bofetada.
      2. slap
      1. He was compensated / medicated / slapped.
      1. (ii)
      1. *Ele recebeu medo / tristeza / vómitos / bebedeira.
      2. Int. He received fear / sadness / vomiting / drunkenness.
    We thank an anonymous reviewer to pointed us this additional criterion to distinguish transferential complex predicates from causative complex predicates. [^]
  12. Nevertheless, they mention that dar corresponding to the verb to become, in Harley’s (2009, p. 333) proposal, is unspecified for the feature [change]. [^]
  13. This is also inter-linguistically supported. See Alba-Sales (2002, p. 38) for empirical evidence from complex predicates with derived and underived Ns in other Romance languages. [^]
  14. See also Alba-Sales (2002) for the discussion of some light verb combinations in Romance which do not have a corresponding main verb. [^]
  15. Cf. Gonçalves et al. (2010, p. 457) for a similar observation with respect to the light verb fazer ‘to do’ in EP (cf. (i)). See also Leal (2009, p. 63).
      1. (i)
      1. O ladrão fez um assalto em cinco minutos.
      2. the thief made a robbery in five minutes.
    [^]
  16. As far as we know, there is no work focusing on aspectual properties of AVP predicates. We will leave this discussion for further investigation. [^]
  17. See also Acedo-Matellán and Pineda (2019, p. 186) for Basque examples, where the same N can combine with different light verbs, without any difference in meaning.
      1. (i)
      1. (a)
      1. argi egin/eman “do/give light” ‘light up, enlighten’
      1.  
      1. (b)
      1. dei egin/eman “do/give call” ‘call’
      1.  
      1. (c)
      1. min egin/eman “do/give pain” ‘hurt’
    [^]
  18. The idea that psychological verbs (of the family of piacere in Italian, as in Il film piace a Maria ‘the movie interests Mary’) are unaccusative was proposed by Belletti and Rizzi (1988), based on properties of these verbs (mainly the fact that their subject can be a Theme and their object an Experiencer). However, as shown by Pesetsky (1995), among others, this sort of verbs does not justify the choice of the auxiliary essere in Italian in the so-called compound tenses (only avere), it does not allow ne cliticization, and allows passives. Moreover, the subject of verbs like piacere may be considered not only a Theme, but a Stimulus or a Cause, therefore, an external argument. Also, it seems difficult that a VP internal argument, as under the unaccusativity hypothesis, should be interpreted as a (subject) Cause. Therefore, the proposal that psychological verbs are unaccusative is highly debatable. [^]
  19. We thank an anonymous reviewer for this suggestion. [^]
  20. Remember that arguments that receive the interpretation of Experiencers, as with psychological predicates, can be subjects, objects, and indirect objects, according to the subclass of psychological predicates: (i) JohnEXP loves / admires / respects Mary; (ii) The tempest frightened / worries BillEXP; (iii) This child gives his parentsEXP some worry / enormous joy (Arad, 1998). [^]
  21. In addition to Cuervo, other authors such as Masullo (1992) and Fernández-Soriano (1999) also argue that, in Spanish, dative experiencers exhibit subject properties. [^]
  22. Cuervo’s (200) analysis of datives in ditransitive structures also entails that Recipients of give-type verbs are not internal arguments of the V, since they are applied datives. [^]
  23. The notion of affectedness adopted here is different from that conceived in the possession relation between the Theme (possessum) and the IO (possessor), as it occurs in external possession structures. [^]
  24. This mixed thematic role assignment is not exclusive of light verb structures, since it is also observed with monotransitive verbs whose Theme has a metaphorical interpretation, as in (i).
      1. (i)
      1. Os alunos estudaram a lição.
      2. ‘The students learned the lesson.’ (Raposo, 2013, p. 376)
    According to Raposo (2013, p. 376), the verb estudar ‘to study’ can assign the thematic role of Agent to its external argument, the Subject, since the act of study is controlled by the students, and the thematic role of Patient to the DO, which can only be (metaphorically) interpreted as an affected entity. Simultaneously, the Subject can be an Experiencer, since students are mentally changed by the action expressed by the verb, whereas the DO is a Stimulus, in the sense that it triggers the Experiencer’s reaction. [^]
  25. See Gonçalves, Duarte and Hagemeijer (2024), for an in-depth analysis of pronominal objects in AVPs. [^]
  26. The idea that complex predicates share the syntactic structure with main verbs was already noticed by Bruening (2016), for English, in his proposal that light verbs are just regular verbs. [^]
  27. Following Gonçalves, Duarte and Hagemeijer (2022), we consider that the head of the KP can be spelled out as a (MOP, STP, AP), para (AP, STP) or em (AP). [^]
  28. See Costa (2009) and Brito (2010) for the proposal that, in EP, IO is generated higher than the DO, as well as Gonçalves (2016) and, in particular, Gonçalves, Duarte and Hagemeijer (2022) for further discussion and implementation of the structure of give-type verbs in Portuguese varieties, using a KP projection. [^]
  29. Costa (2008) discusses the possibility that some unaccusative verbs exhibit lack of subject agreement in colloquial registers of Portuguese, as in (i) and (ii) below. However, the causative structures with dar + N of psychological and physical sensations (in the singular or in the plural) are not exclusive of colloquial registers.
      1. (i)
      1. Fechou aquelas fábricas.
      2. closedSG those factories
      3. ‘Those factories closed’
      1. (ii)
      1. Chegou muitas pessoas.
      2. arrivedSG many people
      3. ‘Many people arrived’
    [^]

References

Acedo-Matellán, V., & Pineda, A. (2019). Light verb constructions in Basque and Romance. In A. Berro, F. Beatriz & J. Ortiz de Urbina (Eds.), Basque and Romance. Aligning Grammars (pp. 176–220). Leiden: Brill. DOI:  http://doi.org/10.1163/9789004395398_007

Alba-Sales, J. (2002). Light Verb Constructions in Romance: a syntactic analysis (Unpublished doctoral dissertation). Cornell University, USA.

Alexiadou, A. (2014) Active, middle, and passive: the morpho-syntax of Voice. Catalan Journal of Linguistics, 13, 19–40. https://revistes.uab.cat/catJL/article/view/v13-alexiadou/153-pdf-en. DOI:  http://doi.org/10.5565/rev/catjl.153

Alonso-Ramos, M. (1998). Étude sémantico-syntaxique des constructions à verbe support [Semantic-syntactic study of light verb constructions]. University of Montreal, Canada. https://papyrus.bib.umontreal.ca/xmlui/bitstream/handle/1866/6777/Alonso_Ramos_Margarita_1998_these.pdf?sequence=1&isAllowed=y

Alsina, A. (1996). The Role of Argument Structure in Grammar. Evidence from Romance. CSLI Publications. Stanford: California.

Arad, M. (1998). VP-Structure and the Syntax-Lexicon Interface (Unpublished doctoral dissertation). University College London, United Kingdom. https://discovery.ucl.ac.uk/id/eprint/10105111/1/U643542.pdf

Bacelar do Nascimento, F. (2013) Processos de lexicalização [Lexicalization processes]. In E. P. Raposo et al. (Eds.), Gramática do Português I (pp. 215–246). Lisboa: Fundação Calouste Gulbenkian.

Belletti, A., & Rizzi, L. (1988). Psych-verbs and Theta Theory. Natural Language and Linguistic Theory, 6, 291–352. DOI:  http://doi.org/10.1007/BF00133902

Brito, A. M. (2008). Grammar variation in the expression of arguments: The case of the Portuguese Indirect Object. Phrasis, 2, 31–58.

Brito, A. M. (2009). Construções de objecto indirecto preposicionais e não preposicionais: uma abordagem generativo-constructivista [Prepositional and non-prepositional indirect object constructions: a generative-constructivist approach]. In A. Fiéis & M. A. Coutinho (Eds.), XXIV Encontro Nacional da Associação Portuguesa de Linguística. Textos Seleccionados (pp. 141–159). APL. https://apl.pt/wp-content/uploads/2017/09/9-Brito.pdf

Brito, A. M. (2022). As construções ditransitivas do Português Angolano e do Português Moçambicano revisitadas [Ditransitive constructions of Angolan Portuguese and Mozambican Portuguese revisited]. In M. Vieira & V. Meireles (Eds.), Variação em português e em outras línguas românicas. Variation en portugais et dans d’autres langues romanes (pp. 163–187). Blucher open access. https://openaccess.blucher.com.br/download-pdf/573/23418. DOI:  http://doi.org/10.5151/9786555501292-09

Brito, A. M., & Oliveira, F. (1997). Nominalization, aspect and argument structure. In G. Matos, I. Miguel, I. Duarte & I. Faria (Eds.), Interfaces in Linguistic Theory (pp. 57–80). Lisboa: APL/Colibri.

Bruening, B. (2015). Depictive Secondary Predicates, Light Verb Give, and Theories of Double Object Constructions (paper accompanies poster presented at NELS 45, MIT). https://udel.edu/~bruening/Downloads/DepictivesDOCs2.pdf

Bruening, B. (2016). Light Verbs are Just Regular Verbs. University of Pennsylvania Working Papers in Linguistics, 22(1), 51–60. https://repository.upenn.edu/server/api/core/bitstreams/1788aa97-579b-45bd-a10e-953ddaf77b82/content

Bruening, B. (2020). Idioms, Collocations, and Structure: Syntactic Constraints on Conventionalized Expressions. Natural Language and Linguistic Theory, 38, 365–424. DOI:  http://doi.org/10.1007/s11049-019-09451-0

Burzio, L. (1986). Italian Syntax. A Government-Binding Approach. Dordrecht: Reidel. DOI:  http://doi.org/10.1007/978-94-009-4522-7

Butt, M. (1995). The Structure of Complex Predicates in Urdu. Stanford: CSLI Publications.

Butt, M. (2010). The Light Verb Jungle: Still Hacking Away. In M. Amberber, B. Baker & M. Harvey (Eds.), Complex Predicates: Cross-linguistic Perspectives on Event Structure (pp. 48–78). Cambridge: Cambridge University Press. DOI:  http://doi.org/10.1017/CBO9780511712234.004

Butt, M., & Geuder, W. (2001). On the (semi)lexical status of light verbs. In H. van Riemsdijk & N. Corver (Eds.), Semi-lexical categories: The function of content words and the content of function words (pp. 323–370). Berlin: Mouton de Gruyter. DOI:  http://doi.org/10.1515/9783110874006.323

Choupina, C., & Brito, A. M. (2018). Dar um beijo é a mesma coisa que dar um espirro? Para uma análise sintático-semântica de diferentes valores do verbo leve dar em Português Europeu [Is ‘giving a kiss’ the same thing as ‘giving a sneeze’? Towards a syntactic-semantic analysis of different values of the light verb dar in European Portuguese]. In J. Veloso, J. Guimarães, P. Silvano & R. Sousa-Silva (Eds.), A Linguística em Diálogo. Volume Comemorativo dos 40 anos do Centro de Linguística da Universidade do Porto (pp. 153–176). Porto: CLUP. https://repositorio-aberto.up.pt/bitstream/10216/119712/2/275492.pdf

Costa, J. (2008). Postverbal subjects and agreement in unaccusative contexts in European Portuguese. The Linguistic Review, 18(1), 1–17. DOI:  http://doi.org/10.1515/tlir.18.1.1

Costa, J. (2009). A focus-binding conspiracy. Left-to-right merge, scrambling and binary structure in European Portuguese. In J. van Craenenbroeck, Alternatives to cartography (pp. 87–108). Berlin / New York: De Gruyter. DOI:  http://doi.org/10.1515/9783110217124.87

Cuervo, C. (2008). Two types of (apparently) ditransitive light verb constructions. In K. Arregi, Z. Fagyal, S. Montrul & A. Tremblay (Eds.), Romance Linguistics 2008: Interactions in Romance (pp. 139–156). Amsterdam & Philadelphia: John Benjamins. DOI:  http://doi.org/10.1075/cilt.313.14cue

Duarte, I., & Batoréo, H. (2004). Predicados complexos e ensino do português europeu como língua não-materna [Complex predicates and teaching European Portuguese as a non-native language]. Palavras, 26, 39–44.

Duarte, I., Colaço, M., Gonçalves, A., Mendes, A., & Miguel, M. (2009). Lexical and syntactic properties of complex predicates of the type <light verb+deverbal noun>. Arena Romanistica. Journal of Romance Studies, 4, 48–57.

Fábregas, A., Jiménez-Fernández, A., & Tubino, M. (2017). What’s up with dative experiencers. In R. E. V. Lopes, J. Ornelas de Avelar & S. M. L. Cyrino (Eds.), Romance Languages and Linguistic Theory 12: Selected papers from the 45th Linguistic Symposium on Romance Languages (LSRL), Campinas, Brazil (pp. 29–48). Amsterdam & Philadelphia: John Benjamins. DOI:  http://doi.org/10.1075/rllt.12.03fab

Fernández-Soriano, O. (1999). Two Types of Impersonal Sentences in Spanish: Locative and Dative Subjects. Syntax, 2(2), 101–140. DOI:  http://doi.org/10.1111/1467-9612.00017

Folli, R., & Harley, H. (2007). Causation, obligation and argument structure: On the nature of little v. Linguistic Inquiry, 38(2), 197–238. DOI:  http://doi.org/10.1162/ling.2007.38.2.197

Folli, R., & Harley, H. (2013). The syntax of argument structure: Evidence from Italian complex predicates. Journal of Linguistics, 49, 93–125 DOI:  http://doi.org/10.1017/S0022226712000072

Gonçalves, A., Cunha, L. F., Miguel, M., Silvano, P., & Silva, F. (2010). Propriedades predicativas dos verbos leves: estrutura argumental e eventiva [Predicative properties of light verbs: argument and event structure]. In A. M. Brito, F. Silva, J. Veloso & A. Fiéis (Eds.), Encontro Nacional da Associação Portuguesa de Linguística XXV, 2010. Textos Selecionados (pp. 449–464). Porto: APL. https://repositorio-aberto.up.pt/bitstream/10216/136543/2/501840.pdf

Gonçalves, R. (2016). Construções ditransitivas no português de São Tomé [Ditransitive constructions in Santomean Portuguese]. (Unpublished doctoral dissertation). University of Lisbon, Portugal. https://repositorio.ul.pt/handle/10451/26409?locale=en.

Gonçalves, R., Duarte, I., & Hagemeijer, T. (2022). Dative Microvariation in African Varieties of Portuguese. Journal of Portuguese Linguistics 21, 6, 1–39. DOI:  http://doi.org/10.16995/jpl.8488

Gonçalves, R., Duarte, I., & Hagemeijer, T. (2024). Objetos diretos em variedades africanas do português: Um estudo de caso de microvariação [Direct objects in African varieties of Portuguese: A case study of microvariation]. In S. Brandão & S. Vieira (Eds.), Para o estudo comparativo de variedades do Português: Questões teórico-metodológicas e análises de dados (pp. 53–84). Berlin: Mouton de Gruyter. https://www.degruyter.com/document/isbn/9783110670257/html#APA. DOI:  http://doi.org/10.1515/9783110670257-005

Grimshaw, J. (1990). Argument Structure. Cambridge: The MIT Press.

Grimshaw, J., & Mester. A. (1988). Light Verbs and θ-Marking. Linguistic Inquiry, 19(2), 205–232.

Gutiérrez Ordóñez, S. (1999). Los dativos [Datives]. In I. Bosque & V. Demonte (Eds.), Gramática descriptiva de la lengua española (pp. 1855–1930). Madrid: Colección Nebrija y Bello, Espasa.

Hagemeijer, T., Mendes, A., Gonçalves, R., Cornejo, C., Madureira, R., & Généreux, M. (2022). The PALMA corpora of African varieties of Portuguese. In N. Calzolari, F. Béchet, Ph. Blache, Kh. Choukri, Ch. Cieri, Th. Declerck, S. Goggi, H. Isahara, B. Maegaard, J. Mariani, H. Mazo, J. Odijk & S. Piperidis (Eds.), Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022) (pp. 5047–5053). Paris: European Language Resources Association. www.lrec-conf.org/proceedings/lrec2022/pdf/2022.lrec-1.539.pdf

Hale, K., & Keyser, S. J. (2002). Prolegomenon to a theory of argument structure. Cambridge, Mass.: MIT Press. DOI:  http://doi.org/10.7551/mitpress/5634.001.0001

Huddleston, R., & Pullum, G. (2002). The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. DOI:  http://doi.org/10.1017/9781316423530

Jespersen, O. (1965). A Modern English Grammar on Historical Principles, Part VI. Allen & Unwin.

Jiménez-Fernández, A., & Rozwadowska, B. (2017). On subject properties of datives in psych predicates. A comparative approach. Acta Linguistica Academica, 64(2), 233–256. DOI:  http://doi.org/10.1556/2062.2017.64.2.4

Kratzer, A. (1996). Severing the external argument from the verb. In J. Rooryck & L. Zaring (Eds.), Phrase Structure and the Lexicon (pp. 109–37). Dordrecht: Kluwer. DOI:  http://doi.org/10.1007/978-94-015-8617-7_5

Leal, A. (2009). Semântica aspectual e nominal. Contributo das expressões nominais para a construção aspectual das frases [Aspectual and nominal semantics. The contribution of nominal expressions to the aspectual construction of sentences]. (Unpublished doctoral dissertation). University of Porto. Portugal.

Marantz, A. (1993) Implications of Asymmetries in Double Object Constructions. In S. A. Mchombo (Eds.), Theoretical aspects of Bantu Grammar, vol. 1 (pp. 113–150). Stanford: CSLI Publications.

Masullo, P. J. (1992). Incorporation and Case Theory in Spanish: a cross-linguistic perspective. (Unpublished doctoral dissertation). University of Washington. USA.

Moens, M. (1987). Tense, Aspect and Temporal Reference. (Unpublished doctoral dissertation). Edinburgh University. Scotland.

Moens, M., & Steedman, M. (1988). Temporal Ontology and Temporal Reference. Computational Linguistics, 14, 15–28.

Mohanan, T. (1994). Argument Structure in Hindi. Stanford, CA: CSLI Publications.

Newman, J. (1996). Give: A Cognitive Linguistic Study. Berlin: Mouton de Gruyter. DOI:  http://doi.org/10.1515/9783110823714

Pesetsky, D. (1995) Zero Syntax: Experiencers and Cascades. Cambridge, Mass: The MIT Press.

Pylkkänen, L. (2002). Introducing arguments (Unpublished doctoral dissertation). MIT, USA.

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

Raposo, E. P. (2013). Estrutura da Frase [Sentence Structure]. In E. P. Raposo et al. (Eds.), Gramática do Português, vol. I (pp. 303–398). Lisboa. Fundação Calouste Gulbenkian.

Rappaport-Hovav, M., & Levin, B. (2008). The English dative alternation: The case for verb sensitivity. Journal of Linguistics, 44, 129–167. DOI:  http://doi.org/10.1017/S0022226707004975

Rassi, A., & Vale, O. (2013). Tipologia das construções verbais em Português do Brasil: Uma proposta de classificação do verbo dar [Typology of verb constructions in Brazilian Portuguese: A proposal for classifying the verb dar]. Caligrama, 18(2), 105–130. DOI:  http://doi.org/10.17851/2238-3824.18.2.105-130

Rio-Torto, G., Soares Rodrigues, A., Pereira, I., Pereira, R., & Ribeiro, S. (2013). Gramática Derivacional do Português. Coimbra: Imprensa da Universidade de Coimbra. DOI:  http://doi.org/10.14195/978-989-26-0641-5

Rosen, S. T. (1989). Argument Structure and Complex Predicates. (Unpublished doctoral dissertation). Brandeis University, USA.

Rothstein, S. (2004). Structuring events. A study in the semantics of lexical aspect. Oxford: Blackwell. DOI:  http://doi.org/10.1002/9780470759127

Scher, A. P. (2004). As construções com o verbo leve dar e nominalizações em –ada no português do Brasil [Constructions with the light verb dar and nominalisations in -ada in Brazilian Portuguese]. (Unpublished doctoral dissertation). University of Campinas, Brazil.

Sleeman, P., & Brito, A. M. (2010). Nominalization, Event, Aspect, and Argument Structure: A Syntactic approach. In M. Duguine, S. Huidobro & N. Madariaga (Eds.), Argument Structure from a Crosslinguistic Perspective (pp. 113–129). Amsterdam & Philadelphia: John Benjamins. DOI:  http://doi.org/10.1075/la.158.07sle

Viñas-de-Puig, R. (2008). Agentivity and experiencer verbs in Catalan and Mayangna and the roles of ‘little v’. Journal of Portuguese Linguistics, 7(2), 151–172. DOI:  http://doi.org/10.5334/jpl.132

Viñas-de-Puig, R. (2014). Predicados psicológicos y estruturas con verbo ligero: Del estado al evento. [Psych predicates and light verb constructions: From state to event]. Revista de Lingüística Teórica y Aplicada, 52(2), 165–188. DOI:  http://doi.org/10.4067/S0718-48832014000200008

Wittenberg, E. (2016). With Light Verb Constructions from Syntax to Concepts. Potsdam Cognitive Science Series: Universitätsverlag Potsdam. https://publishup.uni-potsdam.de/opus4-ubp/frontdoor/deliver/index/docId/8236/file/pcss07.pdf

Wittenberg, E., Jackendoff, R., Kuperberg, G., Paczynski, M., Snedeker, J., Wiese, H., & Wittenberg, E. (2014). The processing and representation of light verb constructions. In A. Bachrach, I. Roy & L. Stockall (Eds.), Structuring the Argument: Multidisciplinary research on verb argument structure (pp. 61–80). John Benjamins Publishing Company. DOI:  http://doi.org/10.1075/lfab.10.04wit