Journal of Hispanic Research 1 (1992-93), pp.167-181

Aspect and Voice: Questions about passivization in Spanish

CHRISTOPHER J. POUNTAIN

Queens' College, Cambridge

1 Introduction

1.1 The Spanish exponents of the passive voice that I intend to consider in this article are what may be conveniently known as the ser-passive (SP), the estar-passive (EP) and the reflexive passive (RP).[1] In some ways this list is arbitrary. It is certainly not exhaustive, there being many aspectually marked periphrastic passives in Spanish formed by verbs of motion (ir, andar, seguir, quedar, venir) with the past participle (pp) (see Green 1982). I understand the distinctive properties of the passive, which these three constructions have in common, to be as follows:

(a) The direct object of the corresponding active verb acts as a grammatical subject. For this reason, the so-called 'impersonal reflexive' of Spanish, which is often listed among translations of the English passive in pedagogical grammars of Spanish (see, for example, Ramsden 1959: 244-46), eg se adora a los héroes, is excluded from consideration, even though the non-specification of a subject (see below) is a property shared with passive constructions. The lack of number agreement between adorar and héroes, and the presence of the 'personal' a, shows that héroes is the object rather than the subject of the verb.

(b) The subject of the corresponding active verb may appear in the passive in an agentive phrase, though this property is constrained, and some purists deny this possibility for RP (Agencia Efe, 1986: 48).[2] The absence of an agentive phrase in a passive construction is not, however, a grave disadvantage. It has been clearly demonstrated (Green 1982 and references there) that the majority of actually occurring passives in Spanish have no agent expressed, and the vagueness achieved by non-expression of the agent must be one of the chief pragmatic motivations for the use of the passive.

1.2 The constraints on these three passives in Spanish have been only very partially dealt with on the whole - perhaps not surprisingly, since it is clear that a variety of factors must simultaneously be taken into account. Three such factors are:

(a) Restrictions of a lexical nature, which presumably account for the unacceptability of

(1) *El cuarto estaba limpiado (cf GRAE: 419)

versus the acceptability of

(2) La calle estaba alargada

I have been unable to discover any convincing reason why limpiar does not have an EP. Native speakers often refer to the existence of the adjective limpio, which combines with estar to give the reading 'to be clean'; but this, whilst close to the projected meaning of the inadmissible *estar limpiado 'to be (in a state of having been) cleaned', would not be synonymous with it, as the possibility for distinction in the English glosses shows. My provisional conclusion, therefore, has to be that limpiar is idiosyncratically constrained.

(b) Constructional homonymy in surface patterns, which presumably explains the non-availability of a SP for aburrir 'to bore':

(3) *Los estudiantes fueron aburridos por la clase

ser aburrido being preferentially construed as copula + adjective 'to be boring'.

(c) Register dependency (demonstrated in Green 1975).

1.3 I suggest that constraints on the Spanish passives can only be understood by appeal to a range of such factors. My purpose here is to look at another of these constraining factors, the aspectual properties of the Spanish passives, but in so doing I do not expect that aspectual properties, any more than any other one of the factors already mentioned, will provide an exhaustive account of the phenomena.

2 The ser-passive (SP)

2.1 It is well known that SP is heavily constrained in the so-called imperfective tenses of Spanish: the imperfect and the present. One constraint in this area has to do with the interaction between morphological and lexical aspect; a very common assertion, dating back to Bello (see Gili Gaya 1948: 124; Fente 1971; 182), is that a verb stem of perfective lexical aspect used in SP cannot have an overall perfective aspectual reading but yields instead a repetitive aspectual reading. Thus:

(4) La puerta es/era abierta por el portero

must imply repetition (todos los días, cada mañana, etc.); it cannot have the single event imperfective reading 'The porter is/was opening the door'. Actually, the lexical aspectual distinction seems rather to be based on the stative/dynamic opposition as conventionally understood (see Comrie 1976: 48-51, Quirk et al. 1972: 39-40); this may be seen from the fact that 'imperfective' but dynamic verbs like comer, escribir, etc., are subject to the same constraints as abrir:

(5) El helado es/era comido por la niña
'Ice cream is/was (e.g., frequently) eaten by the little girl'

likewise has no 'single event' imperfective reading. In fact, we are led to the conclusion that the active sentences corresponding to (4) and (5) in their 'single event' imperfective reading do not have a passive form at all, since it turns out that neither RP nor EP is available either. It follows too that where such a repetitive aspectual interpretation is unavailable, SP is impossible:

(6) *Un gol era metido por Hugo en el minuto veinte (example from Gómez 1988: 187)

Compare the unacceptability of English 'A goal was (*being) scored by Hugo in the twentieth minute', and the repetitive aspect implicit in

(7) Las llamadas a casa del escritor eran respondidas por un contestador automático (El País Internacional, 20.2.89)
'Calls to the writer's house were being answered by an answering machine'

With verb stems of stative lexical aspect, however, SP is indeed the exact equivalent of the corresponding active sentence:

(8) ...esta cualidad [...] que en medida bastante es poseída por todo pueblo saludable (example from Fernández 1986: 421)

which is equivalent to [...] que posee todo pueblo saludable and

(9) Ese médico es respetado por todos (example from Porroche 1988: 69)

which is equivalent to todos respetan a ese médico.

Confirmatory evidence of the dynamic/stative basis of the restriction is offered by verbs which belong to both categories but for which only a stative interpretation is possible in the present or imperfect SP. Thus:

(10) Las causas del desastre eran sabidas
'The causes of the disaster were (well-)known', not '[...]were getting to be known/were being discovered'

2.2 This situation may, however, be undergoing change.

2.2.1 First, in journalistic Spanish and in historical narrative it is increasingly common to find SP used in the present tense with dynamic lexical stems as a kind of historic present or commentary form:

(11) En aquel momento el gol es anulado por el árbitro sin motivo aparente (example from Gómez 1988: 187)
'At that moment the goal was disallowed by the referee without apparent reason' (The past-referring adverbial clearly shows the 'historic present' usage.)

and

(12) El féretro con los restos de Josu Muguruza es llevado a hombros por sus compañeros y amigos (photograph caption, El País Internacional, 27.11.89)
'The coffin with the remains of Josu Muguruza is carried on the shoulders of his companions and friends' (a commentary on a scene)

A sentence like (12) clearly supplies the missing imperfective passive equivalent noted above, since it corresponds exactly to

(13) Sus compañeros y amigos llevan el féretro[...]

The imperfect may also have a similar stylistic use akin to the 'imparfait de rupture' which is familiar from French (Imbs 1960: 93):

(14) Poco después era conducido ignominiosamente a la prevención (Palacio Valdés, quoted in Fernández 1986: 424)
'Shortly afterwards he was being ignominiously led to the police station'

2.2.2 Secondly, there is a marked tendency for the imperfect SP to be used with the pp of a dynamic verb to yield an overall imperfective aspectual reading which is exactly that of the corresponding active form, thus:

(15) En la repleta tribuna miles de aficionados eran comprimidos por otros miles, que pugnaban por entrar (El País Internacional, 17.4.89: 21)
'On the crowded stand thousands of fans were being crushed by thousands more, who were fighting to get in'
= [...] otros miles de aficionados comprimían a los miles de la repleta tribuna
'[...] thousands more fans were crushing the thousands on the crowded stand'

and

(16) Había bajado a la capital para[...]convencer a sus jefes sobre la necesidad de cambiar de estrategia, porque sus muchachos eran diezmados por el Ejército (Allende, Eva Luna: 202)
'He had come down to the capital to persuade his bosses of the necessity of changing strategy, because his lads were being decimated by the Army'
= [...] el Ejército diezmaba a sus muchachos
'[...] the Army was decimating his lads'

It has sometimes been suggested on the basis of the English translation equivalents in (15) and (16) that SP in the imperfect is an avoidance of the much castigated English calque estar + siendo + pp = 'to be being + pp' (Agencia Efe 1986: 50; Lorenzo 1971: 124). While this may be true of direct translations from English (e.g. in journalistic reports based on English sources; see Kany 1945: 237-38), it is difficult to accept this for the literary use in (16). In any case, what the English past progressive and the Spanish imperfect SP have in common is their representation of an imperfective action rather than a state. The non-progressive forms which might be used in the possible English translations 'thousands of fans were crushed by thousands more' and 'his lads were decimated by the Army' are in fact ambiguous between perfective action and imperfective state (in Spanish terms between 'miles de aficionados fueron/estaban comprimidos por otros miles', 'sus muchachos fueron/estaban diezmados por el Ejército'); imperfective action in the past can in English be rendered unambiguously only by the progressive, whereas Spanish has the imperfect tense available.

2.3 What seems to be happening in modern Spanish, then, is that the lexical/morphological possibilities for SP are being more fully exploited with the result that new overall aspectual readings are being obtained.

3 Reflexive passive (RP)

3.1 RP may be considered aspectually neutral; it combines equally with dynamic and stative stems and perfective and imperfective inflections:

(17) Se venden/vendían/vendieron plátanos en el mercado[3]

and

Se le considera/consideraba/consideró tonto

3.2 The constraints on RP are essentially non-aspectual. They are traditionally stated under two headings: (a) the impossibility of expressing an agent, and (b) the necessity for avoidance of collision of surface forms. To exemplify (a) first:

(18) *La puerta se abrió por el guardián

is unacceptable; if an agent is to be expressed, recourse must be had to SP or to another active construction which will foreground the direct object such as

(19) La puerta la abrió el guardián

Restriction (b) is consequent upon the astonishingly multiplex values of the reflexive in modern Spanish: in principle, a reflexive which can be interpreted literally, mediopassively or intransitively cannot be interpreted as a passive; thus:

(20) Mario se despidió sin motivos (example from Porroche 1988: 68)
'Mario took his leave without reason'

cannot be the equivalent of

(21) Mario fue despedido sin motivos
'Mario was dismissed without reason'

and in

(22) Con un martillazo propinado sobre la bola propia, sujeta por el pie, y que repercutía, por reflejo, en la del vencido, ésta era alejada de su itinerario (example from Fernández 1986: 421)

era alejada, 'was sent off course (by the hammer-blow)' cannot be substituted by se alejaba, which would have the meaning of 'was going away (of its own accord)'.

3.3 The basis of the contrast between RP and SP is often described (Fernández 1986: 419; Butt and Benjamin 1988: 298-302; Fente 1971: 183, and references there; Gómez 1988: 186) as relying on the actual or understood involvement of an agent. When no specific agent can be involved, RP is obligatory:

(23) Se produjeron incidentes (Gómez 1988: 186)

but not

*Fueron producidos incidentes

When an overt agent is involved, RP is strictly impossible because of the constraint on the expression of an agent already noted:

(24) *Se publicará un comunicado por el Gobierno Civil (example from Agencia Efe 1986: 48)

Between these extremes one can only speak of tendencies and preferences. The distinction between a true minimal pair such as

(25) Se ha descubierto / Ha sido descubierto un arsenal de armas en un piso (example from Agencia Efe 1986: 47)

relies on the fact that an agent is implied in SP whereas RP carries no such implication.[4]

3.4 Indeed, in Spanish, examples of the breaking of the constraints on RP identified above are increasingly found.

3.4.1 The expression of an agentive por-phrase has often been explained away by regarding the por-phrase as a manner adverbial, which indeed is quite possible, e.g.:

(26) Se supieron las causas del desastre por María
'The causes of the disaster became known through María'

but there seems little doubt of the full agentivity of the por-phrase in an example like

(27) El más antiguo tratado de trigonometría esférica escrita en el mundo se escribió precisamente por un cadí o un juez de Jaén (RNE 10.84: 4-15)
'The oldest treatise on spherical trigonometry written in the world was indeed written by a cadi or judge from Jaén'

especially since this example, taken from a radio transcript, was corrected by the purist editor in a footnote to fue escrito/lo escribió. I have been unable so far to establish the range of the availability of the agentive por-phrase with the RP; textual examples are uncommon, and informant judgments potentially unreliable because of strong normative pressure. One set of judgments I elicited was:

(28) Este libro se publicó por Longman's
Este libro se escribió por un profesor muy conocido

??La casa se edificó por Gómez
*Este cuadro se pintó por Tàpies

and it may well be that the ability of verbs in RP to take por agents is lexically idiosyncratic. However, the mere possibility of a por-phrase in some cases suggests that RP and SP are increasingly perceived by speakers as being closely related in nature. Furthermore, the equivalence of SP and RP for some speakers is suggested by examples like

(29) Entonces Granada se organizó como ciudad enteramente cristiana, según lo habían sido las otras ciudades de su reino (Menéndez Pidal 1962: 126)
'Then Granada was organized as an entirely Christian city, just as the other cities of his kingdom had been (organized)'

where the lo is presumably the proform for an understood pp organizadas.

The incipient movement towards the acceptance of agentive phrases with RP would not be surprising. Even Fernández, who makes a detailed case for semantic distinction between the RP and the SP, has to conclude that in some cases there is near synonymy:

(30) Según elegante frase del historiador Floro, que se había grabado en un pedestal de mármol (= había sido grabada) (example from Fernández, 1986: 420)

3.4.2 The restriction on collision of surface reflexive structures may also be undergoing change. The sentence

(31) La actriz, que llegó a fotografiarse ataviada con un casco norvietnamita durante la guerra... (El País Internacional, 20.6.88)
'The actress, who was even photographed / managed to get herself photographed adorned in a North Vietnamese helmet during the war...'

does not have the expected literal reflexive reading 'even photographed herself' but rather the passive reading 'was even photographed'. Matarse now seems quite firmly established in the meaning of 'to get killed' as well as 'to kill oneself'.[5]

3.4.3 One is tempted to see the striking productivity of the prefix auto- in modern Spanish (see Lang 1990, 182 and 233) as a corollary of this process, foregrounding as it does, redundantly from the etymological point of view (autosuicidarse 'to commit autosuicide (on oneself)' is a case in point) the genuine reflexive nature of the verb:

(32) Un año más tarde se autoexilia definitivamente de Colombia (El País Internacional, 5.3.90)
'A year later he finally went into voluntary exile (not "was exiled") from Colombia'

The signs are, then, that RP may be in the process of completing its trajectory towards a passive which is aspectually neutral with regard to dynamic/stative lexical aspect, even at the expense of former semantic and syntactic constraints.

4 The estar-passive (EP)

4.1 The aspectual contrast between EP and SP is very well-known: SP foregrounds action, whereas EP expresses state. Both Gili Gaya (1948, 110) and Alonso (1951, 242, n.1) are adamant that there is always such a difference. Alonso quite rightly points out the difficulty foreigners have in appreciating the difference. The inadequacy of English in rendering the subtlety of a sentence like

(33) ...que toda la iglesia está llamada a vivir en comunión con Dios y es enviada por Cristo para salvar el mundo (ABC, 19.10.87)
'that all the church is called (has been called?) to live in communion with God and is sent (is being sent?) by Christ to save the world'

is patent. Despite this, it is undeniably the case that the close association between a resultant state and the action which produces that state means that with dynamic verbs EP is often used in a sense which is almost exactly that of the perfect SP haber sido + pp:

(34) Los bolsos están [= han sido] diseñados para proporcionar una mayor comodidad en los viajes (advert in Tiempo, July 1984)
'The bags are (have been) designed to afford greater comfort on journeys'

Furthermore, even Gili Gaya (1948, 110) speaks of a possible overlap in the perfective tenses, between, for example:

(35) Las casas fueron edificadas con mucho cuidado

and

Las casas estuvieron edificadas con mucho cuidado

4.2 There are also aspectually governed constraints on the use of EP.

4.2.1 What we might describe as the 'regular' situation is that an agent cannot be expressed with EP when the verb-stem is inherently perfective in lexical aspect, that is to say, where the pp expresses the resultant state of an action which has terminated (and hence in which the involvement of the agent has also terminated):

(36) La puerta estaba abierta (*por el guardián)

4.2.2 Also 'regular' is the possibility of expressing an agent with EP when the verb-stem is imperfective in lexical aspect and has the property that the resultant state is necessarily contemporaneous with the action, i.e. the action, and hence the involvement of an agent, must continue for the state to exist:

(37) España está representada por el vicepresidente del Gobierno

Notice that the lexical aspectual basis for this distinction is perfective/imperfective rather than dynamic/stative. The categories have not been distinguished satisfactorily in Spanish literature I have encountered on the subject, where dynamic tends to be equated with perfective and stative with imperfective. But many verbs which admit an agent with EP are indubitably dynamic: bloquear, demostrar, mandar, gobernar, provocar, reconocer, to name but a few. Gómez (1988, 189) makes the interesting suggestion that a sentence like

(38) El garaje está vigilado por el guarda

is parallel to the progressive active

(39) El guarda está vigilando el garaje

rather than to the aspectually unmarked active

(40) El guarda vigila el garaje

4.2.3 It will be remembered that the combination with the progressive is viewed as the acid test of dynamicity in English (Quirk et al 1972, 94-7); if this test turns out to be appropriate for Spanish too, it suggests that vigilar is dynamic. Imperfective but stative verbs such as saber, conocer, do not admit EP at all:

(41) *El secreto está sabido

4.2.3.1 However, the relation between stativeness and imperfectivity is close; I have been unable to find for Spanish any candidates for the category of stative perfective verb, which suggests that all Spanish statives may be redundantly imperfective. An interesting collision is hence theoretically possible in Spanish when we have a dynamic imperfective verb yielding in the present tense an imperfective SP and an imperfective EP as in

(42) El mundo es/está gobernado por Dios
'The world is governed by God'

Strictly, the difference must be that SP represents the imperfective action and EP the imperfective ongoing state in which the agent is still involved. However, this subtlety appears to elude even native speakers. Porroche (1988, 74) cites this example as one in which 'apenas existen diferencias'. Compare also the apparent free variation in:

(43) Aquella historia del golpe frustrado parecía poner de relieve que los instrumentos de poder no eran controlados por éste... Allí donde el poder de la policía es incontrolado... ¿Está controlado el poder de nuestra policía por un Gobierno que tiene que dar su palabra de honor... de que no ha espiado a los partidos políticos? (Tiempo, July 1984) [my underlining]

4.2.3.2 Another factor which may be drawing SP and EP more closely together is that EP appears to be extending its domain in modern Spanish, in journalistic register especially, in an extremely interesting way. In the sentences

(44) La obra a que nos referimos está acabada en 1951 y publicada al año siguiente (Lorenzo 1971, 85)
'The work to which we refer was finished in 1951 and published the following year'

(45) El nuevo convenio...deberá estar aprobado por las Cortes antes del próximo día 14 de mayo (El País, 31.10.87)
'The new agreement must have been approved by the Cortes before 14 May next'

we find EP apparently used to represent an action rather than a state. In (44), the present tense cooccurring with a clear past-referring adverbial suggests literal equivalence with the perfective fue (ha sido?) acabada. This usage is sometimes thought of as being restricted to certain registers, but I have been surprised at its ready acceptance, within register limits, by native speakers. Furthermore, it is readily accommodated in the aspectual scheme of Spanish (we might gloss (45) as 'must have gained the approval of the Cortes'), and it can only be the closeness of the association between EP representing resultant state and the corresponding perfective SP representing the action producing that state that has brought about this development.

5 The modern Spanish situation in its diachronic context

These various observations on the current situation in Spanish suggest a diachronic development which I will now sketch.

5.1 It is our experience of Latin (and, for English speakers at least, perhaps of English too) that hoodwinks us into thinking that SP is somehow the archetypal and most neutral form of the passive voice in Spanish. I suggest that in fact there is no archetypal passive in Spanish: as we have seen, SP, EP and RP all have particular limitations. It is well-known that EP is an aspectually marked form; its ability to take agentive phrases is constrained by what I have demonstrated to be consequences of its aspectual markedness. But SP is actually aspectually marked too, in the sense that there is a constrained relation in SP between morphological and lexical aspect. RP is aspectually neutral, but there are general semantic constraints on its use with animate subjects and a prohibition on the expression of an agentive phrase. Moreover, SP and RP, even when agentless, are not generally synonymous.

5.2 SP and EP may therefore be seen today as being the most versatile and frequently occurring members of a whole paradigm of aspectually marked periphrastic passive forms. The aspectually marked nature of SP should not surprise us in view of what we know of its formal history. A dynamic perfective such as CANT(AT)UM EST originally had the perfective meaning 'it has been sung, it was sung' and even when EST was reconstructed as a present it presumably went through a stage of having the perfective meaning 'it is having-been-sung' rather than the imperfective meaning 'it is being-sung'. A stative imperfective such as SAPUTUM EST, on the other hand, with the reconstructed meaning 'it is having-been-known', more easily extended to the present stative imperfective passive represented today by es sabido. The potential final step, still not taken in Spanish, in the aspectual generalisation of es cantado was presumably retarded by the growing availability of the aspectually neutral RP. So much appears to be shared to a greater or lesser degree by all the Romance languages, Spanish RP having gone further than French (Lyons 1982, 179) though not as far as Romanian, where the reflexive readily admits an agent, and where the a fi + pp passive has only latterly been 'restored' in the language.

5.3 Medieval Spanish, on the other hand, sees the encroachment of estar in the copular area and the crucial creation of a surface opposition between ser/estar + pp (Pountain 1982, 157-8). The continual reinforcing of that opposition by the extension of the combination of estar with a very wide range of pps comes to channel out (or rather make overt) an aspectual contrast.

5.4 So far, the picture is of the development of three distinct constructions. But the reflexive and the estar + pp construction share with the so-called 'passive' (ie SP) the feature that their syntactic subject is in fact the semantic object of their verb; the aspectual constraints on SP make it advantageous for RP and EP (and indeed other aspectually marked verb + pp constructions) to supplete SP and indeed to rival or exceed it in frequency.

5.5 The distinctness of these three constructions would be reduced if the following things happened:

1 EP could occur with all or most of the pps with which SP can occur

2 EP were to admit an agentive phrase

3 RP were to admit an agentive phrase

4 The semantic constraint on RP subjects were abandoned

5 SP were to extend its aspectual range

I have presented evidence that stages 1 and 2 are already well under way and that stages 3,4 and 5 are now sporadically occurring. It is perhaps not too far-fetched to speculate that such a convergence, if it leads to greater interchangeability, may eventually challenge the aspectual basis of the contrast among the three passive forms under discussion, or lead to a differently based contrast among them. However, I think that so far the cases of near synonymy among SP, EP and RP that I have identified in modern Spanish are still quite consistent with the aspectual contrast and do not actually threaten it to an important degree. Furthermore, such ongoing convergence as there may be is clearly long-term, and therefore meanwhile Spanish remains endowed with a comparatively subtle aspectual system in the area of the passive.

Notes

1 The material in this article was first presented as a paper at the Nineteenth Romance Linguistics Seminar, Trinity Hall, Cambridge, January 1991. I am grateful to Wendy Bennett, John Green and Chris Lyons for their helpful comments on that occasion.

2 But see the more 'permissive' examples in GRAE, 379.

3 The Spanish distinction between imperfect and preterite is impossible to render in English without elaborate paraphrase. The imperfect has the meaning of 'were being sold'; the preterite represents the action as a complete event, or as an event within a sequence of events.

4 Here, incidentally, may be an answer to the interesting question posed by Lyons 1982, 178-9, as to why middle se in French, the formal cognate of Spanish RP, is excluded from punctual (ie perfective) tenses: perfective tenses are more likely to resist RP since they usually represent singular actions which implicitly involve an agent and favour SP. The rule relating SP, RP and the potential expression of an agent is presumably more clearcut in French than in Spanish.

5 See Pountain 1985, 354 n.6. Interestingly, matar has an 'alternative' suppletive past participle, muerto, in literary usage (Seco 1986, 255-6).

References

Agencia Efe, 1986. Manual de español urgente, 3rd ed. (Madrid:Cátedra).

Alonso, Amado, 1951. 'Sobre métodos: construcciones con verbos de movimiento en español', in Estudios lingüísticos: temas españoles, pp.230-87. (Madrid: Gredos).

Allende, Isabel, 1987. Eva Luna (Barcelona: Plaza y Janés).

Bello, Andrés & Cuervo, Rufino J., 1960. Gramática de la lengua castellana, 6th ed. (Buenos Aires: Sopena)

Butt, John & Benjamin, Carmen, 1988. A New Reference Grammar of Modern Spanish (London: Arnold).

Comrie, Bernard, 1976. Aspect (Cambridge:CUP).

Fente Gómez, Rafael, 1971. Estilística del verbo en inglés y en español (Madrid: SGEL).

Fernández Ramírez, Salvador, 1986. Gramática española: 4 El verbo y la oración (Madrid: Arco).

Gili Gaya, Samuel, 1948. Curso superior de sintaxis española, 2nd ed. (Barcelona: SPES).

Gómez Torrego, Leonardo, 1988. Perífrasis verbales (Madrid: Arco).

GRAE = Real Academia Española, 1931. Gramática de la lengua española (Madrid: Espasa-Calpe).

Green, John N., 1975. 'On the frequency of passive constructions in Modern Spanish', Bulletin of Hispanic Studies, 52, 345-62.

Green, John N., 1982. 'The status of the Romance Auxiliaries of Voice', in Vincent & Harris 1982, pp.97-138.

Imbs, P., 1960. L'emploi des temps verbaux en français moderne (Paris: Klinksieck).

Kany, Charles E., 1951. American-Spanish Syntax, 2nd ed. (Chicago: University of Chicago Press).

Lang, Mervyn F., 1990. Spanish word formation: productive derivational morphology in the modern lexis (London: Croom Helm).

Lorenzo, E., 1966. El español de hoy, lengua en ebullicón, 2nd ed. (Madrid: Gredos).

Lyons, Christopher, 1982. 'Pronominal Voice in French', in Vincent & Harris 1982, pp.161-84.

Porroche Ballesteros, M., 1988. Ser, estar y verbos de cambio (Madrid: Arco).

Pountain, Christopher, '*ESSERE/STARE as a Romance phenomenon', in Vincent & Harris 1982, pp.139-60.

Pountain, Christopher J., 1985. 'Copulas, verbs of possession and auxiliaries in Old Spanish: the evidence for structurally interdependent changes', Bulletin of Hispanic Studies, 62, 337-55.

Quirk, R., Greenbaum, S., Leech, G. & Svartvik, J., 1972. A Grammar of contemporary English (London: Longman).

Ramsden, H., 1959. An Essential Course in Modern Spanish (London: Harrap).

RNE = Radio Nacional de España, Programa de Cooperación Cultural, Madrid.

Seco, M., 1986. Diccionario de dudas y dificultades de la lengua española, 9th ed. (Madrid: Espasa-Calpe).

Vincent, Nigel & Harris, Martin (eds.), 1982. Studies in the Romance Verb (London: Croom Helm).