RRL, LXIII, 3, p. 253–270, Bucureşti, 2018
CONSTRUCTIONS WITH GERUND IN ROMANCE.
A CONTRASTIVE ANALYSIS
*JAN CASALICCHIO
1, LAURA MIGLIORI
2Abstract. Progressive and predicative constructions formed with gerund are present in numerous Romance languages. In the literature, these two types of construction have often been considered as syntactically analogous. Through the application of syntactic tests, this study will show that, despite a number of similarities, progressive and predicative constructs with gerund are characterized by a different structure.
Keywords: gerunds, periphrastic forms, secondary predicates, small clauses, infinitive, non-finite verb forms, monoclausal constructions, biclausal constructions.
1. INTRODUCTION
Numerous Romance languages exhibit a progressive periphrasis formed by an auxiliary and a gerund, as exemplified in (1) and (2):
(1) Estoy estudiando [Spanish]
stay-PRES.1.SG study-GER.
(2) Sto studiando [Italian]
stay-PRES.1.SG study-GER.
‘I am studying
This construct is interpreted with aspectual progressive and/or iterative/continuous reading. Numerous Romance languages also use gerunds in predicative constructions (i.e.
in a secondary predication (3–4)). Most notably, the distribution of progressive and predicative gerunds is not overlapping: in Italian, for instance, gerunds can be used in progressive (2), but not in predicative constructions (5):
* We would like to thank the audience of the 16th Conference on Romance Linguistics, of the Linguistics Seminars of the University Ca' Foscari (Venice) and of the workshop “Participles: Form, Use and Meaning” (SLE Annual Meeting, Zurich 2017) for their useful feedback and suggestions.
Moreover, we would like to thank the editors and the anonymous reviewers for their comments. All shortcomings are of course our own. Casalicchio has received funding by European Union’s Horizon 2020 research and innovation programme under grant agreement no. 681959 (Microcontact).
1 University of Konstanz/Utrecht University, [email protected].
2 Leiden University, [email protected].
(3) Veo a María riendo [Spanish]
see-1.SG to Mary laugh-GER.
‘I see Mary laughing’
(4) Hem vist París amb el Pau fent de guía [Catalan]
have-1.PL seen-PP Paris with the Pau make-GER of guide
‘We saw Paris and Pau was our guide there’
(5) *Vedo Maria mangiando [Italian]
see-1.SG Maria eat-GER.
In the literature to date, gerundial structures in progressive and predicative constructions are often referred to as analogous. More specifically, progressive gerunds are termed as predicative gerunds tout court (cf. Goidanich 1919, Egerland 2010, a.o.). This claim could go back to the hypothesis that progressive gerunds derive from a reanalysis of predicative gerunds, when they were used with verbs of state or movement, such as be, stay, go and come (cf. Corti 1951). An argument in favour of this statement is the presence of gerundial predicative constructions in older stages of those languages that nowadays have only progressive gerunds, e.g. Italo-romance varieties. In many Medieval Italian varieties we find, in fact, instances of predicative gerunds:
(6) […] un bello scudo e di molto valore | a beautiful shield and of much value nel qual vedeasi Marsia sonando (Boccaccio, Tes. 11,6) in-the which saw-SI Marsia playing-music
‘a beautiful and very precious shield, in which Marsia was portrayed playing music’
(7) viti Nicolò d'Autin tegnandose ali caveli see-PAST.1SG Nicolò of Altino keeping-SI at-the hair cum Pelegrin predito (Lio Mazor, 47,9) with Pellegrin aforementioned
‘I saw Nicolò of Altino and the aforementioned Pellegrino tearing each other’s hair out.’
The examples in (6) and (7) come from two different Old Italian varieties: the first is taken from the Florentine writer G. Boccaccio (1339-1341 ca.), while (7) comes from a testimony reported in the archives of Lio Mazor, a town (now disappeared) in the Venetian Lagoon. The documentation of these archives is dated 1312. In both examples the perception verb vedere (‘see’) takes a predicative gerund as its complement, whose subject is a proper noun.
Predicative and progressive gerunds have been considered as related to each other at the synchronic level as well. Raposo’s analysis (1989) proposes that predicative and progressive complements should be analysed as having a similar structure, as shown below3:
(8) [VP Vi [PP os meninosi [P' a [CP/IP eci trabalhar(em)]]]]
3 Note that in predicative constructions both inflected and uninflected infinitives can be used, (8), while in progressive constructions only uninflected infinitives (9). We will come back to this difference infra.
(9) [IP Os meninosi [VP estar [PP os meninosi [P' a [IP os meninosi fumar]]]]]
(Raposo 1989: 295, 298) Moreover, the comparison between these two types of complement is also justified by the possibility of replacing a predicative construction (10a) with a pseudo-relative clause that embeds a progressive form (10b):4
(10) a. Hay dos jovenes fumando [Spanish]
there-are two young smoke-GER. b. Hay dos jovenes que están fumando
there-are two young that stay-3.PL smoke-GER
‘There are two young guys smoking’
(adapted from Fernández Lagunilla 2011) The examples in (10) are existential/locative constructions headed by the form hay (‘there is/are’). In (10), the secondary predicate is formed with a gerund, in (10) with a pseudo-relative clause that hosts a progressive form.
On the basis of these facts, this study aims at investigate whether progressive and predicative constructions with gerund can be really considered as equivalent from a syntactic point of view. This will be done comparing the properties exhibited by these two constructions5.
This paper is structured as follows. In section 2, progressive constructions in Romance and their characteristics will be presented. Then, an overview of Romance predicative costructions will be provided with. After presenting some literature concerning these constructions, the analysis of Casalicchio (2016) will be outlined (section 3). In section 4, a comparison will be made between the properties exhibited by progressive and predicative constructions to assess whether they can be considered syntactically equivalent. Section 5 sketches some preliminary hypotheses concerning the structure of progressive gerunds. Finally, some conclusions and suggestions for future research will be drawn in the last section.
2. GERUNDIAL PROGRESSIVE CONSTRUCTIONS: SYNCHRONIC VARIATION AND PROPERTIES
2.1. Distribution of progressive constructions in Romance
Most Romance languages exhibit a progressive periphrasis formed by an auxiliary (Aux) + gerund. The auxiliary can be stay, as in (11–14), go, as in (15), be (16), or come (17):
(11) Sto andando a casa [Italian]
stay-PRES.1.SG go-GER. to home
‘I am going home’
4 For Pseudo-relative clauses see Cinque (1992) and Casalicchio (2013, 2016), a.o.
5 This work will be concerned with the non-finite verb of progressive constructions, while the issue of the nature and structural position of the auxiliary verb will not be taken into consideration (see Cinque 2006 a. o.).
(12) Estoy comiendo [Spanish]
stay-1.SG eat-GER.
‘I am eating’
(13) Aquêle menino està sempre brigando that boy stay-3.SG always fight- GER. com os outros [Br. Port.]
with the other-PL.
‘That boy is always fighting with the others’
(14) Jo estava parlant amb l’Enric [Catalan]
I-1.SG stay-1.SG talk- GER. with the Enric
‘I was talking to Enric’ (Wheeler, Yates and Dols 1999:365)
(15) ke va truvannə? [Neapolitan]
what go-PRES.3.SG look for-GER.
‘What is he/she looking for?’ (Thomas 1969: 201) (16) so andende [Sardinian]
be-1.sg go-GER.
‘I am going’
(17) Te lo vengo diciendo 2.sg-DAT. it-ACC come-1.SG.PRES. say-GER.
desde el verano pasado [Spanish]
since the summer last (Squartini 1998: 293)
‘I have been saying (lit. come saying) it to you since last summer’
The interpretation of this construct is aspectual progressive/durative. Nonetheless, Romance languages exhibit variation as far as the specific value is concerned, which also depends on the auxiliary verb (cf. Squartini 1998, Vincent 2016). Consider, for instance, the following contrast in standard Italian; while (18a) confers an intensive interpretation (in the sense of Squartini 1998), (18b) gives a purely progressive reading:
(18) a. Cosa vai facendo? [Italian]
what go-2.sg do-GER.
‘What do you keep on doing?’
b. Cosa stai facendo?
what stay-2.SG do-GER.
‘What are you doing (now)?’
A similar contrast can also be observed in Spanish, whereby the stay-periphrasis expresses the pure progressive reading, while the construction with venir has a particular temporal interpretation. More specifically, the latter denotes a durative situation temporally oriented with respect to Reference Time (cf. Squartini 1998):
(19) a. El tren está saliendo [Spanish]
the train stay-3.SG leave-GER.
‘The train is leaving’
b. De cuándo acá vienes preparándome
Since when here come-2.SG-PRES. prepare-GER.-1.SG-DAT. este golpe? (Squartini 1998: 293) this blow
‘Since when have you been preparing this blow?’
In some languages, a prospective interpretation is available as well:
(20) Dile ke esto tornando manyana tell-him-DAT that stay-1.SG.PRES. come-back-GER. tomorrow
‘Tell him that I am returning tomorrow’ [Judaeo-Sp.]
(Malinowsky 1979, in Squartini 1998: 99) Romanian6 and French used to display the progressive construction with gerund at an earlier stage of the language, but they do no longer exhibit it in the modern variety (cf. Werner 1980, Squartini 1998).
(21) a. Il est encore cerchant la sienne 3.SG be-3.SG still look for-PRES.PART./GER. the his
‘He is still looking for his’ [Old French]
b. Il est en train de chercher la sienne 3.SG be-PRES.3.SG in process of look-for-INF. the his
‘He is looking for his’ [Modern French]
(22) a. bărbaţii ceia ( . . . ), sta
men.DEF. NOM those stay.IMPF.3PL
ciudindu-se [Old Romanian]
wonder.GER-CL.REFL.ACC.3.PL
‘And those men were staying wondering’ (Panǎ Dindelegan 2016: 282) b. El lucrează
3.SG work-3.SG
‘He is working’
Furthermore, there are also varieties which exhibit a number of different strategies to express the progressive reading. Consider, for instance, the case of Sicilian, in which a periphrasis with Aux + gerund/a + infinitive is available, (see 23 and 24), next to a construct with BE + an embedded clause, as in (25) (data from Amenta 2010):
(23) a. Sta vivennu [Sicilian]
stay-3.sg live-GER.
‘He/she is living’
b. va camminannu
go-3.sg walk-GER.
‘He/she is walking’
(24) chi ci staju a fari fora?
what LOC. stay-1.SG to do outside
‘What am I doing outside?’
6 With the exception of some Daco-Romance varieties (cf. Marin 1985).
(25) era ca manciava be-3.SG.IMPF COMPL. eat-3.SG.IMPF
‘He/she was eating’
In some varieties, the construct does not exhibit a gerund, but the preposition a + infinitive:
(26) a. ke ʃta a ffa? [Castro dei Volsci]
what stay-2.SG to do-INF. ‘What are you doing?’
b. ʃtɔ a maɲ'ɲa
stay-1.SG to eat-INF.
‘I am eating’
Also in this case, a prospective interpretation is possible, next to the progressive one, as exemplified below:
(27) a. ke 'stai a ffa? [Romanesco]
what stay-2.SG.PRES. to do-INF.
‘What are you doing?’
b. 'stai a vve'ni do'mani 'sera?
stay-2.SG.PRES. to come-INF. tomorrow evening
‘Are you coming tomorrow night?’
Finally, a few Romance languages employ other strategies for expressing the progressive interpretation; generally they use a locative periphrasis (‘be behind’, ‘be in the process/work of…’). This is the case, for instance, of French (see 21b), of numerous Northern Italian dialects (22), and of Rhaeto-romance (28):
(28) So drio magnare [Paduan]
be-1.SG after eat-INF.
‘I am eating’
(29) Son tl lëur de maië [Gardenese (Rhaeto-Romance)]
be-1.SG in-the work of eat ‘I am eating’
The data shown in this section illustrate that there is a certain variation in Romance as far as progressive constructions are concerned. Nevertheless, the construct with Aux + gerund is the most widespread; in fact it is also attested at older stages of languages which do not longer exhibit it.
2.2. Properties of progressive gerunds
The main property of the progressive gerund constructions is the continuous interpretation. This characteristic is present in all the languages displaying this periphrasis
and with all types of auxiliary, but much variation exists concerning the exact interpretation of the construct (cf. Vincent 2016). There is a number of different nuances associated to Aux + gerund, which vary depending on the auxiliary used, but also cross-linguistically.
The most generic/typical reading is purely progressive and refers to an action taking place at the reference time (see Giorgi and Pianesi 1997, Manzini, Lorusso and Savoia 2017). In some languages, like standard Italian, this construct is incompatible with a perfective interpretation:
(30) a. sto andando [Italian]
stay-1.SG go-GER. ‘I am going’
b. stavo andando
stay-IMPF.1.SG go-GER.
‘I was going’
c. *sono stato / fui andando
be-1.SG stay-PP be-PAST-1.SG go-GER.
In other languages, however, the continuous reading does not exclude the combination with a perfective interpretation, like in Sardinian, (31) and in standard Spanish (32):
(31) so istatu travallande [Sardinian]
be-1.SG stay-PP work-GER.
‘I have been working’ (Jones 1993: 141) (32) Estuve esperando en el aeropuerto
stay-1.SG.PAST wait-GER. in the airport más de cuatro horas
more than four hours
‘I was waiting at the airport for more than 4 hours’
(Kattàn-Ibarra et al. 1997: 96) Secondly, the Aktionsart of the verbs used in progressive constructions: activities ((33)a and (34)a) and accomplishments ((33) and (34)) are ruled in, while achievements are very marginal (33) and states are generally ungrammatical (33) and (34c)7:
(33) a. stai lavorando da tre ore [Italian]
stay-2.SG work-GER. since three hours
b. stai leggendo il libro da tre ore stay-2.SG read-GER. the book since three hours
7 For the terminology, see Vendler (1957). Notice, however, that progressives are fine with states in numerous Southern Italian Dialects (cf. Amenta 2010, a.o.):
(i) doppu tṛ̣i uri ca stava stannu aggritta after three hours that stay-1.sg-impf. stay-GER. standing
mi fìciru tràsiri [Sicilian]
1.SG-ACC. let-PAST.3.SG enter-INF.
‘After three hours that I was standing, they let me enter’ (Amenta 2010: 7)
c. ??stai trovando le chiavi stay-2.SG find-GER. the keys d. *stai essendo stanco
stay-2.SG be-GER. tired
e. *stai assomigliando a tua madre stay-2.SG know-GER. at your mother
(34) a. Estuvimos haciendo ejercicio por tres horas [Spanish]
stay-1.PL.PAST do-GER. exercise for three hours ‘We were exercising for three hours’
b. Estuvimos cocinando el conejo durante tres stay-1.PL.PAST cook-GER. the rabbit in three horas
hours
‘We were cooking the rabbit in three hours’
c. *Está sabiendo inglés
stay-3.SG know-GER. English
In this sense, these gerunds sharply differ from adverbial gerunds, which are perfectly fine with states:
(35) Essendo alto, può prendere la scatola lì sopra be-GER. tall can-3.SG take-INF. the box there up ‘Being tall, he can take the box up there’ [Italian]
Another characteristic of progressive constructions is the fact that the gerund has no independent Tense, as shown in (36). This is another crucial difference with adverbial gerunds, in which Tense is independent:
(36) *Oggi sto mangiando ieri today stay-1.SG eat-GER. yesterday
(37) Mangiando così tanto (adesso), stasera non eat-GER. so much (now) this.evening not
avrai fame
have-2.SG.FUT. hunger
‘Eating so much (now), you won’t be hungry this evening.’
To sum up, the main properties of progressive constructions with a gerund are: (i) the continuous/durative reading, (ii) incompatibility with states, (iii) tense dependency8.
8 Another observation regards constituency. In progressive constructions, the gerund forms a constituent on its own, as it can sometimes occur in isolation:
(i) Cosa sta facendo Maria? Mangiando what STAY-3.SG do-GER. Mary eat-GER.
‘What is Mary doing? Eating’
(ii) E' MANGIANDO che sta, non bevendo
be-3.SG eat-GER. that stay-3.SG not drink-GER.
‘He/she is eating, not drinking’
3. GERUNDIAL PREDICATIVE CONSTRUCTIONS: SYNCHRONIC VARIATION AND PROPERTIES
3.1. Distribution of predicative gerunds in Romance
Predicative gerunds are used in several Romance languages, such as Brazilian Portuguese, Spanish, Catalan, Romanian.9 Predicative gerunds can be used in a series of contexts, such as depictive predicates (i.e. adjuncts, (38)), argumental secondary predicates (39), and noun-modifiers (40):
(38) a. Vi a Jorge comiéndose la manzana [Spanish]
see-1.SG.PAST to George eat-GER.-SE the apple ‘I saw George eating an apple’
b. Llegó a casa oliendo a vino arrive-3.SG.PAST to home smelling to wine
‘He reached home smelling of wine’
(39) Sorprendí a Pablo robando catch-1.SG.PAST to Pablo steal-GER. ‘I caught Pablo stealing’
(40) La foto de Ana sonriendo es preciosa the picture of Ana smile-GER. is beautiful
‘The picture in which Ana smiles is beautiful’
The gerundial constructions in (38) are depictive, because they act as adjuncts: the sentence would be grammatical even without them. On the other hand, in (39) a secondary predicate is mandatorily required by the verb sorprender, which is a three-places predicate when it has the meaning ‘catch someone doing something’. Finally, in (40) the gerund is inserted in a DP, modifying it: Ana is represented in this picture in a specific state, namely smiling.
In the past, predicative gerunds were more widespread throughout the Romance world: we find them also in Old Northern Italian and in Old Portuguese, where they were later replaced by prepositional infinitives (see Casalicchio 2013, 2016b, 2017 for Northern Italy, Cunha 1986 and Afonso Pereira 2015 for European Portuguese). Recall examples (6) and (7), repeated here:
(41) […] un bello scudo e di molto valore | a beautiful shield and of much value nel qual vedeasi Marsia sonando (Boccaccio, Tes. 11,6) in-the which saw-SI Marsia play-music-GER.
‘a beautiful and very precious shield, in which Marsia was portrayed playing music’
Notice, however that speakers’ judgement are not homogeneous in this respect. Moreover, in the case of (ii) the absence of the auxiliary could just be a phonological deletion, which does not prevent the Aux to be present (even though silent) in the syntactic structure. Therefore, we will not consider the data above as relevant for our analysis.
9 French shows an idiosyncratic situation, because it uses the so-called participe présent, which is probably the result of the crossing of the latin participium and gerundium (Iliescu and Mourin 1992). Therefore, we do not consider French here.
(42) viti Nicolò d'Autin tegnandose ali caveli see-PAST.1SG Nicolò of Altino keeping-SI at-the hair cum Pelegrin predito (Lio Mazor, 47,9)
with Pellegrin aforementioned
‘I saw Nicolò of Altino and the aforementioned Pellegrino tearing each other’s hair out.’
For sake of clarity, in this article we limit the discussion mostly on predicative gerunds selected by perception verbs.
Some of the languages that do not use a predicative gerund resort to a prepositional infinitive: this holds e.g. for European Portuguese, Gallo-Italic varieties and Standard Italian (with some restrictions, see Casalicchio 2016a, 2017):10
(43) a. e l’æ višt Giórz a mangé ‘r mæ
1.SG.CLCL.have see-PP Giorgio to eat-INF. the apple ‘I saw Giorgio eating the apple’ [Viola, Piedmontese]
b. Chegou a casa a cheirar a vinho [Eur. Port.]
arrive-3.SG.PAST to home to smell-INF. to wine ‘He reached home smelling of wine’
(44) Ho sorpreso Paolo a rubare [Italian]
have-1.SG caught-PP Paolo to steal-INF.
‘I have caught Paul stealing’
3.2. Properties of predicative gerunds shared with progressive gerunds
The properties of progressive gerunds listed in §2.2 hold also for predicative gerunds (Di Tullio 1998, Fernández Lagunilla 1999, Casalicchio 2013)11:
10 Perceptive constructions with bare infinitives (i), on the other hand, have a different structure, as shown e.g. by the fact that they do not need an overt semantic subject, unlike predicative gerunds (ii):
(i) Vi a María comer/ comiendo
see-PAST.1.SG. to Maria eat-INF. eat-GER. ‘I saw Maria eat.’
(ii) Oí __ gritar / *gritando
hear-PAST.1.SG cry-INF. cry-GER. ‘I heard someone cry.’
11 Note that the properties listed in this section, which mainly come from Di Tullio (1998) and Fernández Lagunilla (1999), all hold for predicative gerunds used in perceptive constructions. When they are used in different contexts, some of these restrictions do not hold: predicative gerunds are also possible with states (i), see Fernández Lagunilla (2011); and they can have independent tense (ii):
(i) Fabio llegó de/a Londres sabiendo inglés Fabio arrived from/to London knowing English Fabio knew English when he arrived from/to London
(Fernández Lagunilla 2011, ex. (15a)) (ii) Con María llorando toda la noche, mañana estaré muy cansado with Maria crying all the night tomorrow be-FUT.very tired ‘If Maria keeps crying all the night, tomorrow I will be very tired.’
(i) Predicative gerunds have continuous/progressive aspect, as noted by Di Tullio (1998).
Consider a sentence like:
(45) Vi el barco atracando en el puerto see-PAST.1SG the ship dock-GER. in the port
‘I saw the ship docking at the port’
This does not imply that the ship docked indeed at the port: this example can be followed by a sentence saying “but it went adrift because a hurricane came up”. Thus, the perception focuses on a single moment within the process of docking, but this does not mean that the event came to its natural end.
(ii) Fernández Lagunilla (1999) reports that predicative gerunds can only be built with accomplishments and activities, but neither with states nor with achievements:
(46) Vi a María trabajando / leyendo un libro / see-1SG.PAST to Maria work-GER. read-GER. a book
*encontrando las llaves / *sabiendo inglés find-GER. the keys know-GER. English
As shown in (46), predicative gerunds of work (activity) and read a book (accomplishment) are fine, but those of find the keys (achievement) and know English (state) are not.
(iii) Predicative gerunds have dependent/anaphoric tense, like progressive gerunds:
(47) Ayer vi a Pablo corriendo (*anteayer / *hoy / *mañana) yesterday see-1SG.PAST to Pablo run-GER.
(the.day.before.yesterday/today/tomorrow)
In (47), the reference time of the matrix verb see and of the gerundial verb run must be partially overlapping, in the sense that the time of the event of seeing must be contained in the time of the event of running. Since the perception took place yesterday, the event of running cannot have taken place on a different moment.
4. GERUNDIAL PROGRESSIVES AND PREDICATIVES: THE SAME STRUCTURE?
In the previous paragraphs (§ 2. and 3.) we have shown that progressive and predicative gerunds share a number of properties. However, here we illustrate that there are some fundamental differences between the two constructions, concerning both the size of the gerundial complement and its relationship with the matrix verb.
The first two differences concern negation. First of all, the event expressed by the gerund can be negated, using the construction ‘sin (‘without’) + infinitive’, in the case of predicative (48), but not of progressive gerunds (49):
(48) a. Nunca te he visto escuchando música never you have-1SG see-PP listen-GER music b. Nunca te he visto sin escuchar música [Amer. Spanish]
never you have-1SG. see-PP without listen.INF music
(49) a. Estoy comiendo [Spanish]
stay-1.SG eat-GER. b. #Estoy sin comer stay-1.SG whithout eat-INF.
In (48), the event expressed by the gerund escuchando música (‘listening to music’) can be negated by a prepositional infinitive headed by ‘without’. This possibility is not given in progressive gerunds, where the embedded event cannot be negated.
Second, when the matrix event is negated, negation has scope over a progressive, but not over a predicative gerund:
(50) No está comiendo (= she didn't eat at all) not stay-3.SG eat-GER.
‘She is not eating’
(51) No la vi comiendo (= she may have been eating or not) not her see-1.SG-PAST eat-GER.
‘I did not see her eating’
Third, the unmarked position of the semantic subject of the gerund is different: in Spanish, for instance, it is between the matrix verb and the gerund in the case of predicative gerunds (52), while it precedes the auxiliary when a progressive gerunds is used (see 53):
(52) Veo a Pablo comiendo (unmarked)
see-1.sg to Pablo eat-GER.
‘I see Pablo eating’
(53) a. Pablo está comiendo (unmarked)
Pablo stay-3.SG eat-GER.
b. Está Pablo comiendo (marked) stay-3.SG Pablo eat-GER.
‘Pablo is eating’
Note that (53) is even ungrammatical in other Romance languages like Italian, where (53a) is the only grammatical order.
Furthermore, when a prepositional infinitive is used in a European Portuguese progressive construction, the infinitive must be uninflected. In predicative constructions, on the other hand, it can be either inflected or uninflected (Raposo 1989, Duarte 2003)12: (54) Vi os meninos a devorar(em) o gelado
see-1.sg.past the children to eat(INFL.) the ice-cream
‘I saw the children eating the ice-cream’
12 Recall also footnote 3.
(55) Os meninos estão a devorar(*em) o gelado the children stay-3.PL to eat(*INFL.) the ice-cream
‘The children are eating the ice-cream’ (Duarte 2003: 643)
While the contrast between (54) and (55) does not directly affect gerunds, because EP always uses prepositional infinitives in progressive and predicative constructions, it is nonetheless telling for the size of the two constructions: predicative constructions seem to have more structure, since they allow also for inflection, while progressive constructions are more reduced.
Moreover, predicative gerunds can be analysed as Small Clauses (in the sense of Williams 1975, Moro 1993, 1995, a.o.). As shown in Casalicchio (2013, 2017), they have the same distribution as adjectival and prepositional Small Clauses (56), and they can be coordinated with them 0. This property does not hold for progressive gerunds.
(56) a. Vi a Juan muy preocupado (AP) [Spanish]
see-1.SG.PAST to Juan very worried
de muy buen humor (PP)
of very good humour
nadando (Pred. Gerund)
swimming
‘I saw Juan very worried/in a wonderful mood/swimming’
b. Vi a María en la cama see-1.SG.PAST to Maria in the bed
y a Antonio viendo la televisión [Spanish]
and to Antonio watch-GER. the televisión
‘I saw Maria in bed and Antonio watching TV’ (Casalicchio 2017)
(57) a. Gianni sta *malato (AP) [Italian]
Gianni stay-3.SG ill
*di buon umore (PP) of good mood
nuotando (Progr. gerund)
swim-GER.
b. *María está en su cama y Ana cocinando
Maria stay-3.SG in her bed and Ana cook-GER. c. *Pablo está en el sofá y comiendo Pablo stay-3.SG in the couch and eat-GER. Finally, if progressive and predicative gerunds were the same, we would expect their cross-linguistic distribution to overlap, but as we have seen this is not the case: in some languages only progressive, in others only predicative gerunds are used. Even the fact that languages like Italian had predicative gerunds in past stages cannot be considered as evidence for the same structure, because when Italian predicative gerunds evolved into prepositional infinitives, progressive gerunds should have shared the same evolution.
Table 1 summarises the shared properties and the differences between the two gerundial constructions13.
Table I – Summary
Progressive cstr. Predicative cstr.
Tense is anaphoric + +
Progressive aspect + +
Not stative + +
Negation with sin – +
Negation scope over the gerund not over the gerund
Extraction + Constrained
Grammatical in Italian yes no
European Portuguese inflectional
infinitives – +
Subject position before the gerund – (or marked) +
To sum up, even though progressive and predicative gerunds share a number of properties, they display significant differences as far as their syntactic behaviour is concerned. Therefore, we conclude that they are characterised by a different underlying structure.
5. TOWARDS AN ANALYSIS OF PROGRESSIVE GERUNDS
Offering a detailed analysis of progressive gerunds is beyond the aims of this paper.
In this section, we would like to offer some preliminary observations about the structure of progressive gerunds, contrasted with that of predicative gerunds, with which they share some properties, although the structure is not the same.
13 One additional difference between predicative and progressive gerunds concerns extraction, although things are not completely clear. Extraction out of a progressive gerund is always grammatical (i); with predicative gerunds, extraction is only possible when the semantic subject of the gerund is in post-verbal position, but not when it is in the unmarked, preverbal position (cf. (52)), witness (ii):
(i) a. ¿Qué está comiendo Pablo?
what stay-3.SG. eat-GER. Pablo
b. ¿Qué estará Pablo comiendo?
what stay-3.SG.FUT Pablo eat-GER. 'What is Pablo eating?'
(ii) a. *¿Qué viste leyendo a Juan?
what see-2.SG.PAST read-GER. to Juan
b. ¿Qué (le) viste a Juan leyendo?
what (him) see-2.SG.PAST to Juan read-GER. 'What did you see Juan reading?'
First of all, according to the literature progressive and predicative gerunds share the presence of a locative preposition. For progressive gerunds, this has been proposed e.g. by Mateu (2002), who also builds on Bolinger’s (1971) analysis of English gerunds. Both authors convincingly show that progressive gerunds are related to a locative PP, as can be seen in the couples of expressions below:
(58) a. He is at work He is working b. She is at prayer She is praying
(Bolinger 1971: 247 ff., cited in Mateu 2002: 137) (59) a. de fiesta festejando
of party celebrate-GER.
b. de caza cazando (Masullo 2008, cited in Gallego 2010: 88) of hunt hunt-GER.
Mateu (2002) proposes that the gerund incorporates a preposition of central coincidence (Hale 1985), which is responsible for the locative/progressive reading of the gerund: the preposition marks the relation between the embedded verb and the subject, which is centrally located in the event expressed by the gerund. An analysis along similar lines has also been put forward by Fábregas (2008), Gallego (2010), Gallego and Hernánz (2012), and Fábregas and Jiménez-Fernández (2016), a.o. For predicative gerunds, on the other hand, they have been analysed as incorporations of a preposition in Casalicchio (2013, 2017), who based his proposal on the comparison with prepositional infinitives.
Most recently, Silvagni (2017) has proposed that Spanish gerunds have a more complex structure: they are not a PP, but a Predicative Phrase that hosts a PP. This PredP gives them a stage-level nature which makes it possible for them to be selected by estar.14 On the other hand, Manzini, Lorusso and Savoia (2017) have claimed that the preposition a found in Southern Italian progressive constructions is the same preposition used to mark dative/locative. They analyse this dative/locative/progressive preposition as marking a part/whole, or inclusion relation (⊆): crucially, in progressive constructions it instantiates not the relation of the embedded event with the subject, as in Mateu’s (2000) work, but with the matrix event. In fact, they consider progressive constructions as biclausal: the matrix and the embedded event are both merged in two separate VPs. Casalicchio (2017) proposes to extend the presence of a preposition of inclusion relation to predicative gerunds.
In this paper, we limit ourselves to highlight some points that are useful for an analysis of progressive gerunds, and also for a formal differentiation between progressive and predicative gerunds. The literature to date, cited in this section, shows that both types of gerund can be considered as the result of an incorporated preposition. Nonetheless, as we have shown above, these constructions crucially differ as far as their syntactic structure is concerned15. Predicative gerunds are Small Clauses, while progressive gerunds are not. For the latter, as a first approximation, we propose that they are mono-clausal structures, with
14 Silvagni’s (2017) dissertation is mainly focussed on a formal explanation of the selection process of the copulas ser and estar in Spanish. He considers that estar has a interpretable [Stage]
feature which matches an uninterpretable [Stage] feature present in stage-level predicates.
15 Interestingly, progressive and predicative gerunds may differ also with respect to their size, i.e. progressive gerunds may have a more reduced TP (or they may even be just a vP) than predicative gerunds, which have at least the aspectual projections of the TP (Casalicchio 2017). We leave this issue open for future research.
the embedded verb merged in VP and then moved to an aspectual projection within TP, and the auxiliary stare merged in a higher functional projection of the same TP, following Cinque’s (2006) analysis of functional heads (see also Grano 2015):
(60) [FPJohn try [vP John to open the door]] (Grano 2015: 5) (61) [FP Maria sta [vP Maria lavando i panni]]
This allows us to straightforwardly account for the scope of the matrix negation over the event expressed by the gerund and for the unconstrained extraction out of a progressive gerund.
Finally, consider the examples below:
(62) a. La pietra sta rotolando [Italian]
the stone stay-3.SG roll-GER. ‘The stone is rolling.’
b. *La pietra sta mangiando
the stone stay-3.SG eating-GER.
These facts can be accounted for under a mono-clausal analysis, which explains the selectional restrictions on the subject: the matrix subject must be compatible with the requirements of the embedded verb and not of the auxiliary16.
A final observation in favour of a mono-clausal analysis is given by the possibility of clitic climbing:
(63) a. Sto ascoltando la musica
stay-1.SG listen-GER. the music ‘I am listening to music’
b. La sto ascoltando
her-CL stay-1.SG listen-GER. ‘I am listening to it’
In light of these properties, our preliminary hypothesis is that gerundial progressive constructions are characterized by a mono-clausal structure.
6. CONCLUSIONS AND HYPOTHESES FOR FUTURE RESEARCH
In this paper we have compared progressive and predicative constructions with gerund. We have shown that the properties shared by both constructs do not provide us with conclusive evidence in favour of an unified analysis. These characteristics can be better considered as typical of gerunds in general and not really construction-specific (cf. also Lonzi 1991 for Italian gerunds).
16 For a different proposal that analyses progressive constructions as biclausal, see Manzini, Lorusso and Savoia (2017), see supra.
Conversely, it has been illustrated that progressives and predicatives exhibit crucial differences as far as their syntactic behaviour is concerned, which indicates that they have a different underlying structure. We consider predicative constructions to be Small Clauses, while progressives display different properties. A preliminary analysis of these periphrases suggests that gerundial progressives are syntactically mono-clausal.
REFERENCES
Afonso Pereira, S., 2015, “Predicative constructions with gerunds in European Portuguese dialects”, Dialectologia. Special Issue, 5, 351–371.
Amenta, L., 2010, “Perifrasi verbali in siciliano”, in: J. Garzonio (ed.) Studi sui dialetti della Sicilia.
Special Issue of Quaderni di lavoro ASIt, 11, 1–18.
Bertinetto, P. M., 2000, “The progressive in Romance, as compared with English”, in: Ö. Dahl (ed.), Tense and aspect in the languages of Europe, Berlin & New York, Mouton de Gruyter, 559–604.
Bolinger, D., 1971, “The Nominal in the Progressive”, Linguistic Inquiry, 2, 2, 246–250.
Casalicchio, J., 2013, Pseudorelative, gerundi e infiniti nelle varietà romanze. Affinità (solo) superficiali e corrispondenze strutturali, München, Lincom Europa.
Casalicchio, J., 2016a, “The use of gerunds and infinitives in perceptive constructions: the effects of a threefold parametric variation in some Romance varieties”, in: E. Bidese, F. Cognola, M. Moroni (eds), Theoretical Approaches to Linguistic Variation, Amsterdam, John Benjamins, 53–87.
Casalicchio, J., 2016b, “Ricostruire la “diacronia della sintassi ladino-dolomitica con l’aiuto di Joppi.
Il caso dei costrutti percettivi”, in: F. Vicario (ed.), Ad limina Alpium. VI Colloquium Retoromanistich, Cormons, dai 2 ai 4 di Otubar dal 2014, Udine, Società Filologica Friulana, 97–126.
Casalicchio, J., 2017, The syntax of Romance predicative gerunds and prepositional infinitives, Ms., Utrecht University.
Cinque, G., 2004, ““Restructuring” and functional structure”, in: A. Belletti (ed.) Structures and beyond: The Cartography of Syntactic Structures, New York – Oxford, Oxford University Press, 132–191.
Cinque, G., 2006, Restructuring and Functional Heads, Oxford, Oxford University Press.
Corti, M., 1953, “Studi sulla sintassi della lingua poetica avanti lo Stilnovo”, Atti e Memorie dell'Accademia Toscana di Scienze e Lettere La Colombaria, n.s., 4, 261–365.
Cunha, C., 1986, “Conservação e inovação no português do Brasil”, O Eixo e a Roda, 5, 199–232.
Duarte, I., 2003, “Subordinaçâo completiva - as orações completivas”, in M. H. Mira Mateus et al.
(eds), Gramática da Língua Portuguesa, Lisboa, Caminho, 593–652.
Egerland, V., 2010, “Frasi subordinate al gerundio”, in: L. Renzi, G. Salvi (eds), Grammatica dell’italiano antico, vol. II, Bologna, il Mulino, 903–920.
Fernandez Lagunilla, M., 2011, “Aspecto, predicación secundaria y complementos verbales no finitos en español”, in: R. M. Ortiz Ciscomani (ed.), Contribuciones al estudio del español, Universidad de Sonora, Colección Estudios lingüísticos, 4, 15–50.
Goidanich, P.G., 1919, Grammatica Italiana ad uso delle Scuole con nozioni di metrica, esercizi e suggerimenti didattici, Bologna, Zanichelli.
Grano, T., 2015, Control and Restructuring, Oxford, Oxford University Press.
Iliescu, M., L. Mourin., 1992, Typologie de la morphologie verbale romane, Innsbrucker Beiträge zur Kulturwissenschaft, Sonderheft 80.
Jones, M. A., 1993, Sardinian Syntax, Routledge.
Kattàn-Ibarra, J., J. Christopher, 1997, A modern Spanish grammar, London-New York, Routledge.
Lonzi, L., 1991, “Il sintagma avverbiale”, in: L. Renzi (ed.), Grande grammatica italiana di consultazione, vol I, Bologna, Il mulino, 341–412.
Malinowski, A. C., 1979, Aspects of contemporary Judaeo-Spanish in Israel based on oral and written sources, Doctoral dissertation, University of Michigan.
Manzini, M. R., P. Lorusso, L. M. Savoia, 2017, “a/bare finite complements in Southern Italian:
mono-clausal or bi-clausal syntax?”, Quaderni di linguistica e studi orientali, 3, 11–59.
Marin, M., 1985, “Formes verbales périphrastiques de l’indicatif dans les parlers dacoroumains”, Revue roumaine de linguistique, 30, 169–206.
Masullo, P., 2008, “The syntax-lexical interface. Prepositionalizing motion verbs in Spanish”, Ms., University of Pittsburgh.
Mateu, J., 2002, Argument structure: relational construal at the syntax-semantics interface, PhD-dissertation, Universitat Autònoma de Barcelona.
Moro, A., 1993, I predicati nominali e la struttura della frase, Padova, Unipress.
Moro, A., 1995, Small Clauses with predicative nominal, in: A. Cardinaletti, M. T. Guasti (eds), Small Clauses, San Diego, Academic Press, 109–135.
Noetzel, L. M., 2007, “The Spanish Progressive construction and its Latin source”, Neophilologus, 91, 51–61.
Raposo, E., 1989, “Prepositional infinitival constructions in European Portuguese”, in: O. Jaeggli, K.J. Safir (eds), The Null Subject Parameter, Dordrecht, Kluwer, 277–305.
Silvagni, F., 2017, Entre estados y eventos, Unpublished PhD-Dissertation, Universitat Autònoma de Barcelona.
Squartini, M., 1998, Verbal periphrases in Romance. Aspect, actionality and grammaticalization, Berlin, Mouton de Gruyter.
Vendler, Z., 1957, “Verbs and Times”, The Philosophical Review, 66, 2, 143–160.
Vincent, N., 2016, “A structural comparison of Latin and Romance”, in: A. Ledgeway, M. Maiden (eds), The Oxford Guide of the Romance Languages, Oxford, Oxford University Press, 37–49.
Williams, E., 1975, Small Clauses in English, in: J. P. Kimball (ed), Verb Movement. Orlando, Academic Press, 249–273.
