Four Dialectal Uses of the Adverb Siempre and Their Grammatical Properties

Abstract

This article analyzes four interpretations of the adverb siempre ‘always’ that do not belong to general Spanish. The continuative and the progressive-comparative interpretations are argued to be calques of Italian, often attested in Rioplatense Spanish. In the first one, siempre is equivalent to Eng. still or ‘continue to + infinitive’, while in the second one it admits paraphrases with more and more, less and less, and the adverbs gradually and progressively. The third interpretation, in which siempre is roughly equivalent to after all, finally, and ‘end up + gerund’, will be argued to be concessive-adversative. This reading is more frequent in Mexico and Central America, but it is also attested in other American countries. The fourth reading is the attenuated interpretation, registered in part of the Andean area. In this meaning, siempre is equivalent to roughly or so so. It is argued that, with the possible exception of the last reading (whose origin is insecure), these different meanings of siempre coincide in the interpretation of this adverb as a universal quantifier, while they differ in the semantic nature of the quantified variable.

1. Introduction

The present paper addresses the readings of the Spanish adverb siempre ‘always’ that are subject to dialectal variation in the Spanish-speaking world, with special attention to the similarities and differences between them. These uses have been scarcely studied in the literature. In Bosque (2015) both the general, majority, or common uses of this adverb and those interpretations restricted to some geographical areas are addressed. In this paper, I will analyze some similarities between the dialectal (or geographically restricted) usages of siempre and parallel interpretations in the counterparts of this adverb in both general or standard Spanish (in the sense of “the one shared by most speakers”), and other Romance languages. I will also argue that the different senses of siempre (with the possible exception of the last one, whose origin is insecure) coincide in the nature of this adverb as a universal quantifier, while they differ in the quantified entity.

The universal quantification expressed by Sp. siempre is not overt in the etymology of this word, contrary to Fr. toujours (literally, ‘all days’) o Eng. always (that is, ‘all roads, all means’). On the fact that the etymology of Eng. always corresponds to Sp. todavía ‘still’ (toda-vía, lit. ‘all-way’), or It. tuttavia ‘nevertheless’, rather than to that of siempre, see Morera (1999), Giacalone Ramat and Mauri (2009), Bosque (2016), and references therein.

The two fundamental interpretations of the adverb siempre, common to all varieties of Spanish, as well as many other languages, are temporal:

(a)	The first is the durative interpretation. In this reading, siempre quantifies over a set of points in a timeline, as in Kant siempre vivió en Königsberg ‘Kant always lived in Königsberg’. The timeline in this example is provided by Kant’s life.
(b)	In the second interpretation, always quantifies over a set of events, as in Siempre llegas tarde ‘You are always late’. This sentence can be paraphrased as ‘Whenever you arrive, you arrive late’.

The first interpretation requires durative predicates, whereas the second requires events. These may have duration (Juan siempre come aquí ‘Juan always eats here’) or lack it (Siempre llegas tarde ‘You are always late’). As in English, given the appropriate context, with some predicates we can obtain either the durative interpretation (He always read a lot) or the event reading (He always read the newspaper in the morning).

These two interpretations have been extensively studied—as well as related—in the literature and will not be discussed here. Among many other works, the reader my consult von Fintel (1992, 1994), de Swart (1993), Rothstein (1995), Bach et al. (1995), Doetjes (2007), and Lenci and Bertinetto (2008). The fact that sometimes it is not easy to tell these two readings apart might explain why they are often not distinguished in dictionaries as different interpretations. Nowakowska (2010) discusses other consequences of the distinction, as well as its dependency on aspectual factors.

2. The Continuous Interpretation

The first dialectal reading of siempre that I will address is the continuous interpretation. The reading is more properly aspectual than temporal. It admits paraphrases with the adverb todavía ‘still’, and also with the cursive aspectual periphrasis formed by seguir ‘keep’ and a gerund, similar to Eng. continue to plus an infinitive. It is quite natural to suppose that this use of siempre was introduced in the Rioplatense (henceforth, RP) variety as a direct calque from Italian.

Let me exemplify the continuous interpretation of siempre before looking into its properties. The English sentence Do you still live in Rome? corresponds to Italian Vivi sempre a Roma? The equivalent in most varieties of Spanish is ¿Sigues/Seguís viviendo en Roma?, but in RP Spanish one also attests ¿Vivís siempre en Roma? The following example illustrates this pattern in Argentinian Spanish:

(1)	Y que se fijen de mi patrimonio, que vivo siempre en la misma casa y no tengo ni bicicleta (Hebe de Bonafini in Tiempo Argentino, 16 May 2017). ‘And let them look at my assets, I still live in the same house and I don’t even own a bicycle’

This example is particularly shocking to non-RP Spanish speakers, because they tend to interpreted the adverb siempre in the durative reading. This gives rise to the rather absurd interpretation that always provides in this very context in English, namely ‘I live in the same house all the time’, or ‘I live in the same house at all hours’. As is obvious, the correct reading is ‘I still live in the same house’ or ‘I continue to live in the same house’.

The continuous use of siempre is normatively censored for it constitutes a syntactic calque of Italian. The academic DPD illustrates this reading with an example of the Argentinian writer Roberto Cossa:

(2)	¿Recuerdas nuestro bar en la rue la Victoire? [...] ¿Estará siempre Michel? (R. Cossa, El viejo criado, DPD, s/v siempre) ‘Do you remember our bar in Victoria street […] Will Michel still be there?

Again, this sentence is strange for non-RP native speakers of Spanish, as they tend to interpret siempre in the first of the senses described above. This results in the absurd interpretation ‘Will Michel be there all the time?’, instead of the correct reading ‘Will Michel still be there?’. As I have advanced, this reading naturally admits English paraphrases formed with ‘continue to + infinitive’.

Interestingly, the negative version of the latter periphrasis is constructed in French with the adverb toujours ‘always’, something which cannot be done in any varieties of Spanish (with siempre) or English (with always). Therefore, Fr. Marie n’est toujours pas arrivée means ‘Marie still does not arrive’, but it literally says ‘Marie is not always arrived’. On the use of Sp. todavía ‘still’ with the meaning of always in Old Spanish, see Morera (1999), Magaña Juárez (2014), and Bosque (2016).

Before providing other examples of the continuous interpretation of siempre, we may introduce a natural theoretical question: ‘What exactly does the adverb siempre quantify over in this reading (assuming it is still a quantifier)?’. I will argue that siempre remains a universal quantifier in this reading, and also that the quantified variable corresponding to this new usage is represented by a series of time points located in some ongoing succession whose final limit is linked to the moment of speech, or to a certain reference point measured from it. According to this, a possible paraphrase of the sentence Vivo siempre en esa casa (in the now relevant reading ‘I still live in that house’, RP Spanish), would be “For each of the consecutive time points provided by an interval containing the moment of speech, and beginning at a certain past point, the proposition I live in that house is true”.

As can be seen, the entity to be quantified—from this point of view—is the set of time points provided by the continuous and uninterrupted series of stages that continues up to the moment of speech (or up to the point of reference that has been introduced). To this, one must add a modal factor that further links siempre to todavía ‘still’. This is the counterexpectation inference to which this usage of siempre gives rise. It has been extensively studied for still and its counterparts in other languages (see König 1977; Löbner 1989; Mittwoch 1993; Van der Auwera 1993; Michaelis 1996; and Umbach 2012, among many others). The paraphrase that expresses this inference in the example Vivo siempre en esa casa would, approximately, be “It would not be expected that I would live in that house by now”. In more technical terms, the counter-expectation concerns a possible phase transition (ter Meulen 1995; Engerer 2014), since, in unmarked cases, the situation expressed by the predicate is expected to cease from a certain point onwards.

Here are other examples of the continuous interpretation of siempre, all corresponding to RP Spanish. In most of them, but not all, the adverb siempre occupies the post-verbal position, a tendency that was confirmed to me by the speakers of this variety I asked for their opinion on these sentences:

(3)	a. ¿Me querés siempre? ¿No me olvidaste? (L. Ordaz, Breve historia del teatro argentino, gb). ‘Do you still love me? Haven’t you forgotten me?’
	b. ¿Pensás siempre subir al Cerro Empinado? (A. Valessi, El llamamiento, gb). ‘Are you still planning to climb the Steep Hill?’
	c. Alberto trabajaba siempre en el mismo taller, con su padre (E. Friedmann, En busca de un destino, gb). ‘Alberto continued to work in the same workshop with his father’
	d. Pero qué linda oficina que tenés. ¿Siempre conservás la misma fábrica? (N. García Yudé, Pez expreso: cuento y narraciones, gb) ‘What a nice office you have. Do you still have the same factory?’

As it is obvious, the durative reading of siempre is also available in these sentences, but must be contextually discarded. According to this possible (but not relevant) interpretation, (3a) would mean ‘Do you love me always?’, and (3b) would mean ‘Do you think about climbing the Steep Hill all the time?’. These are the only readings that non-RP native speakers of Spanish would attribute to the sentences in (3). In all cases, these interpretations are grammatically possible, but contextually inappropriate.

The RP variety of Spanish is not the only one to behave in this way in the Romania. It is worth remembering that Fr. toujours and It. sempre display a systematic ambiguity between temporal and aspectual readings, the latter corresponding to phasal periphrases in durative contexts. Thus, the French sentence Le concierge est toujours là might mean ‘The concierge is still there’ (so that toujours = still), or rather ‘The concierge is always there’ (so that toujours = always, at all times). In a similar way, to the question—Tu travailles toujours dans le même bureau? ‘Do you still/always work in the same office?’, one might answer—Oui, je n’ai pas changé de travail ‘Yes. I have not changed my job’, so that toujours = still, but it would also be possible to answer—Non, je me déplace parfois ‘No, sometimes I move around’, so that toujours = always. Italian gives rise to a similar ambiguity. As in (3), this possible ambiguity is only resolved depending on the context and the situation.

It is important to emphasize that the continuous use of siempre displays the properties of the aspectual adverb todavía ‘still’ and its close equivalents in Romance and Germanic languages, described in detail in the titles just cited. This does not exactly mean that all the uses that might be covered by the label “continuous” coincide exactly. For example, the use of E(uropean) P(ortuguese) sempre analyzed by Amaral and Del Prete (2014) bears some resemblance with the continuative use of siempre than I am addressing, but, at the same time, there are important differences between them. These authors interpret non-temporal uses of EP sempre as cases of truth persistence, in the sense than they express the overt confirmation of some previously asserted or conveyed content. The EP sentence Sempre vou ao cinema no domingo à noite does not mean ‘I always go to the movies on Sunday night’ (as opposed to Vou sempre ao cinema no domingo à noite), but rather ‘My intention of going to the movies on Sunday night persists or still holds’.

The necessary preverbal position of sempre in the EP sentence sharply contrasts with the postverbal position of RP siempre in the continuative use. Other differences include the fact that EP sempre does not allow for paraphrases with ainda ‘still’ or continuar ‘continue’ plus gerund (even if Eng. still could be appropriate in some of these cases, see below). The association of EP sempre, according to Amaral & Del Prete, with some previous action plan of the speaker is not necessarily found in the RP continuative use of siempre either. In any case, it would be interesting to investigate the points of contact that exist between both usages, as well as the important question of their specific origin. Later on, I will address the issue of whether EP sempre could be more closely related to the use of siempre that I call “concessive-adversative”.

As I have argued, the durative and the continuous interpretation of siempre are clearly different. But there is little doubt that speakers of RP Spanish also speak general Spanish (in the sense of “Spanish shared with other Spanish speakers”). For many speakers of the RP variety, sentences such as Juan piensa siempre en jubilarse are two-way ambiguous. The first of these two readings is the durative interpretation, shared by other Spanish speakers. This reading gives rise to the paraphrase ‘Juan is thinking about retirement all the time’. The second sense is the continuative reading, not shared by most Spanish speakers, which gives rise to the paraphrase ‘Juan is still thinking about retirement. (He hasn’t changed his mind)’.

As I have anticipated, many speakers of Argentinian Spanish only perceive the durative interpretation with pre-posed siempre, as in Juan siempre piensa en jubilarse. One might possibly argue that this is a natural consequence of the well-known fact that temporal projections are located higher than aspectual ones in syntactic configurations. Although this line is worth pursuing, I will not do so in the present paper.

3. The Progressive-Comparative Interpretation

As I pointed out at the outset, siempre may be analyzed as a universal quantifier in most of its interpretations, although it does not quantify over time instances in all readings. In fact, it does not necessarily quantify over them in the interpretation that I will deal with in this section, which is also attested in the RP area as a calque of a common use in Italian.

French and Italian use counterparts of Eng. always in a reading that may be called progressive-comparative. This interpretation, not present in most varieties of Spanish, is obtained in comparations of inequality, and can be paraphrased with the adverbs progressively or gradually, and the (partially iconic) English adverbial expressions more and more and less and less. It also allows for the Spanish paraphrase cada vez (lit. ‘each time’) followed by a comparative phrase. The Spanish sentence El problema se hacía cada vez más complejo ‘The problem was becoming more and more complex’ corresponds to French Le problème devenait toujours plus complexe, and Italian Il problema diventava sempre più complesso. Notice the presence of Fr. toujours and It. sempre (both, ‘always’) in these paraphrases. French can also use the adverb de plus en plus ‘more and more’ instead of toujours plus ‘always more’ in these sentences, while in Italian it is more difficult to avoid the adverb sempre in parallel cases. Note also that the term incremental is not entirely appropriate to describe this usage, since the relevant progression may take place in descending scales, as in It. Le verdure sono sempre meno fresche ‘Vegetables are becoming less and less fresh’. I am assuming that ‘downward incremental scales’ and ‘negative growth’ do not exist, or are to be interpreted as puns.

Some speakers of RP Spanish reproduce this use of It. sempre, which is, again, censured by the Association of Spanish Language Academies (see DPD, p. 602). Here are some examples of this use:

(4)	a. Entre estas libertades, la de prensa asume siempre más importancia (Hoy, 5 November 1986; quoted in DPD, p. 602). ‘Among these liberties, press freedom becomes more and more important’
	b. En general, los fenotipos homocigotos recesivos son siempre menos frecuentes (Agencia de noticias científicas, Universidad Nacional de Quilmes, 28 October 2022) ‘In general, homozygous recessive phenotypes become progressively less frequent’
	c. Carlos tenía la impresión de que los chicos que llegaban al primero de bachillerato eran siempre más pequeños (A. Faccini, El perfume de la adolescencia, gb). ‘Carlos had the impression that the boys who arrived at the first year of high school were younger and younger’
	d. Siempre más triste, siempre más turbio se vuelve el perfeccionamiento de nuestro arte (S. Hahnemann, Escritos médicos menores; translated by F. Darío Flores, gb) ‘The improvement of our art becomes sadder and murkier every day’
	e. También llegan vagones paqueteros, los destinados a encomiendas, que como no llevaban peso eran siempre más chicos (Página 12, 24 June 2002) ‘Parcel wagons […] arrive as well. Since they did not carry weight, they were smaller and smaller’
	f. […] el abandono de los cuarteles se está haciendo siempre más frecuente en las realidades urbanas (F. Camerin, Registros, vol. 14, 2018, p. 163, gb) ‘The abandonment of barracks is becoming more and more frequent in urban realities’
	g. [...] la posesión de computadoras siempre más poderosas y complejas tiene […] un límite preciso (E. Canutti, Ascendentes en Astrología, crea) ‘The possession of ever more powerful and complex computers always has a precise limit’

All the examples in (4), from RP Spanish, contain the adverb siempre, but non-RP speakers of Spanish would not interpret this adverb in the sense provided by the English paraphrases in (4), but rather in the durative reading (‘always’, ‘all the time’, ‘at all times’) mentioned above. The very fact that siempre does not mean ‘always’ but rather ‘progressively’ or ‘gradually’ in these examples characterizes the “progressive-comparative” interpretation. Interestingly, these examples cover almost all registers and styles. Examples (4b,d,g) belong to scientific language; examples (4a,f) are journalistic, and (4c) is from a literary text.

The examples in (4) contain a comparative NP or AP of inequality modified by the adverb siempre. My RP informants tended to reject sentences parallel to the ones in (4) when the comparative quantifier preceded by siempre is adverbial, rather than nominal or adjectival. Therefore, they strongly preferred Trabajás cada vez más to Trabajás siempre más ‘You work harder and harder’ (in the intended reading). Even so, texts in which this usage appears can be found:1

(5)	Aumentan los impuestos, aumentan siempre más, tal vez para Navidades ya no comeremos (El Mundo, 10 November 1996) ‘Taxes are increasing, they are increasing more and more, maybe by Christmas we won’t eat anymore’

Notice that the relevant bracketing in (5) is [aumentan] [siempre más], while in the durative interpretation (non-dialectal and irrelevant here), it would be [aumentan siempre] [más]. Incidentally, the newspaper El Mundo is from Spain, but the journalist who wrote this news item was Argentinian.

I have also encountered speakers of RP Spanish who considered the usages illustrated in (4) to be archaic. In principle, it is not surprising that certain syntactic calques are more prominent among speakers who know the source language (i.e., the language from which the imitated pattern originates), or among first-generation descendants. The fact that these calques do not necessarily extend to descendants of later generations is to some extent to be expected, as a natural consequence of the greater degree of internationalization of Spanish.

Finally, I have found speakers of RP Spanish who accepted the sentences in (4), but rejected them when the series to be created is not strictly temporal, a point in which they clearly deviate from common Italian usages. In these cases, they chose the standard forms cada vez más (‘more and more’; lit. ‘each time more’) or cada vez menos (‘less and less’; lit. ‘each time less’).2 Notice that no event, time point or interval is aligned in Spanish sentences such as those in (6):3

(6)	a. Las ventanas del edificio eran cada vez más chicas. ‘The windows of the building were getting progressively smaller’
	b. Las luces del corredor parecían cada vez más tenues. ‘The lights in the corridor seemed dimmer and dimmer’

Again, the equivalents of (6) in Italian and French contain counterparts of always (namely, It. sempre, and Fr. toujours). As I have indicated, several of my RP informants rejected the counterparts of these sentences with the adverb siempre, while admitting the examples in (4), with temporal uses of siempre in the progressive-comparative interpretation. Others accepted all these examples as natural. I am afraid I have no explanation for this diversity of judgments, although I suspect that age and generation (that is, second, third or fourth generation of native speakers) may be relevant factors.

Let us take a quick look at this use of siempre-cada vez from a theoretical perspective. We may construct approximate paraphrases of Fr. toujours, It. sempre, and (in part) RP Sp. siempre by converting the variables bound by the universal quantifier into a set of nominal entities in an increasing or decreasing succession. This perspective assimilates the present pattern to the one corresponding to sentences in which so-called unselective binding takes place (see Lewis 1975; Heim 1982; Krifka 1988; Chierchia 1995; Carlson and Pelletier 1995; and Dobrovie-Sorin 2003, among many other studies).

Let me clarify this point. Standard illustrations of the unselective binding pattern show that universal quantifiers such as always may pick up non-temporal variables in generic sentences. The adverb always does not exactly mean ‘at all times’ in the English sentence Planets are always bigger than the satellites surrounding them. As a matter of fact, always does not quantify over any possible temporal unit (years, centuries, millennia, eons, etc.) in this sentence. A natural paraphrase of this proposition would be All planets are larger than the satellites surrounding them, a sentence in which some universal quantifier takes planets in its scope. In parallel, always quantifies over vowels (and not over time instances of any kind) in Vowels are always voiced, and many other similar generic sentences. A similar conclusion could arguably be applied to Sp. cada vez ‘every time’, so that the noun vez (‘time instance’, as Fr. fois or It. volta) in the examples in (6) turns out to be a sort of wildcard substitute of the quantified variable. Most probably, all languages with some equivalent of Eng. always allow for this semantic change on the interpretation of the quantified entity. With this in mind, a paraphrase of sentence (6a) would approximately be ‘Given a linear series of windows in some building, each of them was smaller than the one immediately preceding it’.

As I have indicated, only some RP speakers admit siempre instead of cada vez in (6), that is, in cases where the progression that characterizes the relevant series is not temporal. Certainly, the series is temporal in the examples in (4). A simplified version of (4e), from an Argentinian newspaper, is Los vagones eran siempre más chicos ‘Wagons were progressively smaller’. As is obvious, this sentence does not speak of a (somewhat absurd) situation in which a certain train consists of several wagons, each of which is smaller than the wagon preceding it. The sentence speaks, on the contrary, of a certain time series in which there is a progressive reduction in the size of the train wagons. This means that, in such cases, universal quantification does not take place on the individuals forming the series (wagons, in this case), but on the set of time instances at which their size is evaluated.

The basic licensing conditions of the progressive-comparative reading of siempre include a comparison of inequality modified by siempre in a reading close to Eng. progressively. If the series is temporal, all time instances in a contextually bounded period will be relevant; if it is not temporal, all individuals affected by the universal quantifier (windows or lights in the examples above) will be included. Consequently, the universal quantification expressed by siempre applies in both cases.

Although non-selective binding is often associated with universal quantification, it is worth remembering that this is not the only possibility. In Bosque and Masullo (1998) it is argued that mucho ‘much’ is not a frequency adverb in sentences such as La gente compra mucho este libro ‘People buy this book a lot’. If it were, this sentence would mean ‘People buy this book many times’. This interpretation is not correct, since people buy specific books only once (if at all). The sentence proposed really means ‘Many people buy this book’. Again, the quantifier takes people as a variable in this paraphrase, and does not quantity over time units nor events.4

4. The Concessive-Adversative Interpretation

The use of siempre that I now consider is usually paraphrased with adverbs such as definitely, decisively, or finally, but—as I will explain below—it allows for other periphrases which might reflect its meaning more accurately. This use is typical of Mexico, Central America and other countries of the Caribbean area, although Kany (1963, pp. 382–83) also attested it in other areas. It is perceived in sentences such as Siempre no me voy a Europa ‘Definitely, I am not going to Europe’, Siempre se casará el sábado ‘Finally, s/he will marry next Saturday, o ¿Estás decidido a viajar siempre? ‘Are you finally decided to travel?’ According to van Wijk (1969), this usage of siempre is (or was) common in Honduras.

The possible paraphrases of the usage of siempre that I now address do not coincide exactly. Siempre is approximately equivalent to finally in the examples in (7):

(7)	a. Le pedí que se quedara/ que de mí no se alejara/ le rogué que no se fuera/ pero él siempre se marchó (J. Rivera, Se marchó [Mexican popular song]). ‘I asked him to stay/ not to go away from me/ I begged him not to leave/ but he finally left’
	b. ¿Siempre se va mañana? Él sonrió y dijo: —De juro (J. de Viana, Leña seca, from Kany 1963, p. 382). ‘Is s/he finally leaving tomorrow? —He smiled and said —for sure’
	c. [The housekeeper addressing a woman about her baby:] ¿Siempre se murió el tuyo? (M. Azuela, Víctimas de la opulencia y otros relatos, gb). ‘Did yours finally die?’
	d. ¿Se decidió usted a venir siempre? (C. Mangado, Modismos panameños en el lenguaje, from Kany 1963, p. 383). ‘Did you finally decide to come?’

But, in other cases, paraphrases of siempre with definitely, certainly, for sure or for good seem to be more accurate:

(8)	a. Pero a mí me han dicho que tú vas muy seguido a su casa, y eso siempre no me gusta (V. Leñero, Dramas sociales, gb). ‘But I’ve been told that you go to his/her house very often, and I definitely don’t like that’
	b. —Me habían dicho que tú habías resuelto no casarte ya. —Mentira, yo no he dicho nada… —¿Entonces, siempre te casas? (R. Maluenda, Escenas de la vida campesina, from Kany 1963, p. 382). ‘I had been told that you had resolved not to get married anymore. —Bullshit! I didn’t say any of that…—So, are you definitely getting married?’
	c. Ante todo, ¿estás siempre decidido a casarte? (A. Alceda, Vida criolla, from Kany 1963, p. 382). ‘First of all, are you determined to get married for good?’

In the present section I will argue that...

(i)	...the use of siempre analyzed here is concessive-adversative,
(ii)	...this usage involves a form of universal quantification.

Let us consider (i) first. Strictly adversative paraphrases of siempre are provided by Kany in some of his numerous examples attesting this usage. This includes paraphrases with nonetheless, however, and similar expressions in cases like the following:

(9)	a. Aunque la mañana del domingo es lluviosa, siempre se ha levantado mamá (T. Carrasquilla, Tiempos revueltos, from Kany 1963, p. 383) ‘Although Sunday morning is rainy, mom is up nonetheless’
	b. ¿Y aquel tomó? —No, señor. Lo dejó. —Pero lléveselo siempre (A. Ambrogui, El Jetón, from Kany 1963, p. 383). ‘Did s/he take that one from there? —No, sir, He did not. He left it. But you may take it with you nonetheless’

It is worth remembering that the passage from continuous to adversative uses in certain adverbs is a process repeatedly observed in several languages. As a matter of fact, Eng. still sometimes allows for paraphrases with however, nonetheless, nevertheless, though, even so, and similar expressions. The same can be said of Fr. toujours and (more restrictively) of Sp. todavía. The case of It. tuttavia ‘nevertheless’ is even more interesting, since the continuative value of this adverb has given way exclusively to the adversative reading, a process that has been studied in detail in Giacalone Ramat and Mauri (2009) and references therein. The durative (‘always’), continuative (‘still’) and adversative (‘nonetheless’) values of Fr. toujours coexist today in many contexts, but some of these three interpretations of parallel adverbs have been lost in the other Romance languages, or are extremely restricted.

One may now wonder where the quantifying component mentioned in (ii) exactly fits in. There are several arguments in favor of it. Paraphrases with in any case or anyway (corresponding to Sp. en todo caso, en cualquier caso, de todas maneras, de cualquier modo, etc.), are provided by Kany as good equivalents of siempre in some of his examples:

(10)	a. Se despertó y gritaba mucho. Siempre[,] la dejamos (S. Quevedo y Zubieta, La camada, from Kany 1963, p. 383). ‘She woke up and was screaming a lot. In any case, we abandoned it’
	b. —¿Nos vamos o te esperamos?, pues no dilata en reventar la autora. —Siempre vete yendo, compadre (L. G. Inclán, Astucia, from Kany 1963, p. 383). ‘Should we leave or wait for you?, since it won’t be long before dawn. —In any case, get going, my friend’

Notice that the idiomatic expressions just mentioned (en todo caso, de cualquier modo, etc.) coincide in two relevant aspects: (i) they are adversative, and (ii), they contain universal quantifiers. Along similar lines, Curcó (2004) explains that paraphrases of this interpretation of siempre are often obtained with expressions such as in spite of all this or after all. Notice again the crucial presence of the quantifier all in these paraphrases. As in other adversative-concessive expressions, this use of siempre introduces a statement contrary to some content that was previously stated or implied. If we take these considerations into account, we can conclude that, in these cases, the adverb siempre quantifies over the complete (overt or covert) set of discarded options, which gives rise to paraphrases such as ‘Even considering all of the above’. This paraphrase makes it easy to understand examples of Mexican Spanish such as the one quoted in DPD (s/v siempre):

(11)	[Someone opens a tape recorder, takes out the tape and throws it in the trash. His partner replies in astonishment:] ¿Qué, siempre no la vas a usar? (F. Santander, El corrido de los dos hermanos, quoted in DPD, p. 602). ‘What? You’re not going to throw it away after all, are you?

The quantificational value of siempre is therefore maintained in this concessive-adversative interpretation. Notice that siempre is a VP internal adverb in this use, while the expression after all is peripheral. Even so, it is only the antecedent of all (or its covert or deleted complement) that is truly extra-sentential.

The label “concessive-adversative” that I am using reflects that well-known fact that the meaning of some particles (including Ger. doch or Sp. a pesar de) may be paraphrased by concessive connectors (such as although) in some cases, and with adversative particles (such as nonetheless) in some others. Possible paraphrases of siempre with the adverbial expression después de todo ‘after all’ are interesting because they force us to ask ourselves what exactly does todo ‘all’ quantify over in these cases. Interestingly, the presence of this quantifier in similar idiomatic phrases is common in various languages: Eng. all in all; Sp. con todo, con todo y (con) eso (both ‘all in all, yet’), similar to Catalan tot i això ‘although’ (lit. ‘all and that’), It. in tutto ‘all in all’. Sp. con todo y que ‘in spite of’, common in Mexico, is not very far from this use, as in Se le otorgó, con todo y que el permiso anterior tenía todavía vigencia por un año (Proceso [México], 1 December 1996) ‘The license was given to him, although the previous one was still valid for one year’.

Let us now take a look at the paraphrases of siempre with finally, illustrated in (7). In these cases, successive events are not simply aligned, unlike what we might say in He had a wine, a beer and, finally, a whisky, for example. Instead of preceding the last term in some linear succession, siempre introduces in (7) a resolution to be argumentatively evaluated against what can be deduced from the immediately preceding information. As a matter of fact, siempre in (7) allows for paraphrases with both finally and in spite of it all. Similar equivalences are obtained in other cases. As regards paraphrases with adverbs such as definitively, illustrated in (8), it should be noted that in these sentences the context provides information that can be interpreted as argumentatively opposed to the expression introduced by siempre, so that definitively and finally do not express radically different meanings in these cases.

In her analysis of the Mexican uses of siempre, Curcó (2004) provides paraphrases of this adverb with finally, but also with the periphrasis “end up + gerund”. Her example Siempre acepté el trabajo (literally, “I always accepted the job”) is approximately equivalent to Acabé aceptando el trabajo ‘I ended up accepting the job’, or Finalmente, acepté el trabajo ‘Finally, I accepted the job’. This author also provides an explanation of the meaning of this example that makes clear the counterargumentative value of siempre. This sentence…

“…conveys that the speaker accepted a particular job, but also that there has been uncertainty as to whether this would be the case, and that the probability that the speaker would not accept it was high at some stage in the past”.

The Italian and Portuguese equivalents of the cursive periphrases just mentioned and their relationship to sentences with some Romance counterparts of finally are analyzed in detail in Amaral and Del Prete (2016, 2020). These authors argue that these periphrases denote particular forms of non-epistemic modality related to the expectation of the speaker. They interpret them as cases of truth unpersistence, and they argue that they constitute the opposite pattern of the use of EP sempre mentioned in § 2 above; that is, the cases of truth persistence.

As far as my analysis of siempre is concerned, it must be emphasized that it is the same adverb that gives rise to apparently opposite situations (truth persistence vs. truth unpersistence), suggesting that, if one attempts to adapt these analyses to Spanish, at least one of the two interpretations will have to be discarded. I am not opposed to the idea that there might be some relationship between the concessive-adversative reading of siempre and Amaral and Del Prete’s (2014, 2016, 2020) theory on truth (un)persistence, but my analysis of siempre focuses on distinguishing what is shared from what is not shared by the different interpretations of siempre. It is also linked to the much-studied relationship between continuative and adversative uses of the same adverbs in Romance and other languages, as well as the (extensively analyzed) relationship that the former maintain with the concept of “expectation”.

The notion of ‘series’ of ‘succession’ shows that there is a close relationship between the dialectal use of siempre that I am addressing here and a more general interpretation of this adverb that could be dubbed last resource reading. I have in mind a use of siempre in Spanish that has equivalents in most Romance languages (as well as in English), and that admits close paraphrases with the expressions at the very least or ultimately, as in Siempre es una solución ‘It’s always a solution’ (said when referring to a certain possible option, after discarding others mentioned before or deducible from context), or as in Si pierdes el autobús, siempre puedes tomar un taxi ‘If you miss the bus, you can always take a cab’.

The adverb siempre lacks temporal value in this usage, which Cadiot et al. (1985) call “pragmatic”, applied to Fr. toujours, and Mosegaard Hansen (2004) calls “connective”. The term last resource reading does not imply that the option introduced by siempre is extreme. Even so, the most notorious characteristic of this meaning of siempre is its scalar nature, as pointed out by Buchi (2007) in relation to Fr. toujours. Franckel (1989) emphasized the fact that toujours introduces in such cases an alternative hypothesis to those that have been mentioned (and implicitly discarded) in the previous discourse, and Cadiot et al. (1985) pointed out the argumentative nature of this usage, insofar as it is the speaker who tidies up the discursive weight of the options presented, and places one of them in the last place. More precisely, Cadiot et al. (1985) suggest that in these cases Fr. toujours introduces an option that constitutes a ‘weak argument’ (‘neutral’ in Mosegaard Hansen’s 2004 words), which provides a possibly final—and often non-optimal—solution to some situation presented as adverse.

In both the (dialectal) concessive-adversative reading of siempre and the (widespread, standard, or non-dialectal) last resource reading, some final option is presented as not desirable (in principle) or contrary to the natural conclusion that would follow from previous assertions or assumptions. In (7a), several requests are overtly presented as desirable, so that siempre introduces an (undesirable) concluding resolution. In (11), siempre introduces an unexpected alternative (namely, ‘not to use something’), after some previous information that might be expected to give rise to the opposite result. In the (non-dialectal) last resource reading that Spanish shares with other Romance and non-Romance languages, the “last option” introduced by siempre is also presented or offered as not eligible or non-optimal as first choice.

I wish to draw attention to the fact that the idiomatic expression al fin y al cabo ‘in the end, after all’, which provides an approximate equivalent of siempre in the last resource reading, contains the word fin ‘end’, and the word cabo (also ‘end’). Recall that the adverb finalmente ‘finally’ has been mentioned above as one of the possible paraphrases of siempre in the concessive-adversative reading. Expressions such as after all, in the end and finally similarly imply a set of options that ends up being discarded in the presence of a last alternative, deemed feasible.

We may now ask what, then, is the point of contact between the (common or standard) interpretation of siempre that I have called last resource, and the (geographically restricted) reading that I have called concessive-adversative. I would like to suggest that a concessive-adversative meaning is also expressed in the last resource reading, and also that the adverb siempre quantifies over a universal set of unchosen options (whether or not overtly presented) in both cases. The difference essentially concerns the grammatical distribution of this particle—much more restrictive in the last resource reading— rather than the meaning that it contributes.

Let me clarify the latter point. Spanish does not allow for the purely adversative use of siempre (equivalent to Eng. however or nonetheless) that French toujours permits in final position. Thus, if one person assures another in French that he or she has left a certain object in a certain place, the latter could reply in surprise: Je n’ai rien trouvé, toujours ‘But I found nothing’. The literal translation of this sentence into English is ‘I have not found anything, always’. This meaning of toujours is discussed in detail in Nguyen (1988). See also Nguyen (1986a, 1986b) on the relationship between this interpretation and other senses of toujours.

The last resource interpretation of siempre does not save the latter example in English or Spanish, just as it does not save it in imperative contexts. One may say Inténtalo de todas formas ‘Try it anyway’ in all varieties of Spanish, or Inténtalo al menos ‘Try it at least’, but no variety of this language allows for #Inténtalo siempre ‘#Try it always’ with this same meaning (that is, siempre in the sense of Eng. anyway), an option quite natural in French: Essaie toujours (Buchi’s 2007 example).

Although this paper does not deal with the uses of siempre common to all Spanish speakers, it is interesting to note that the (dialectal) concessive-adversative interpretation and the (non-dialectal) last resort reading interpretation are closely related. In Bosque (2015) it is suggested (with a different terminology) that, in the last resource interpretation, siempre could represent a free-choice indefinite, since it admits paraphrases with cualquier(a) ‘any’, en cualquier caso ‘in any case’, de cualquier manera ‘in any way’, etc. It is worth recalling that free-choice items are sometimes paraphrased with universal quantifiers (Any child knows that = All children know that), and with existential quantifiers at other times (Ask anyone ≠ Ask everyone). Negative indefinites, such as Sp. nadie ‘no one, anyone’ are also known to admit both types of paraphrases in the appropriate contexts.

This analysis has positive aspects, but also some negative ones. It could explain the presence of the modal verb poder ‘can, may’ (a well-known intensional trigger of free-choice indefinites) in sentences such as Siempre puedes tomar un taxi ‘You can always take a cab’, in sharp contrast to #Siempre tomé un taxi ‘I always took a cab’, where only the event interpretation is possible. From this point of view, the irregularity of this sentence, in the relevant reading, would be parallel to the anomaly of *Te ha llamado cualquier amigo ‘*Any of your friends called you’, versus Te ha podido llamar cualquier amigo ‘Any friend might have called you’.

Also favoring the analysis of siempre as a free-choice indefinite in the reading above is the fact the future tense (another well-known intensional trigger) facilitates this use, as in Siempre te quedará la satisfacción del deber cumplido ‘The satisfaction of duty fulfilled will always remain in you’. Again, the substitution of quedará ‘will remain’ for quedó ‘remained’ would considerably degrade the grammaticality of the sentence.

On the contrary, against the analysis of siempre as a free-choice indefinite is the fact that the last resort interpretation is rejected with imperatives, as I have just pointed out (a point in which Spanish and French sharply differ), as well as in other typically intensional contexts.5 Even so, the close semantic relationship between the concessive-adversative interpretation of siempre and the last resource reading can be maintained. The former is not restricted to the intensional contexts that (in most cases) characterize the second, but they both introduce similar counter-expected resolutions after mentioning or implying all other options.

5. The Attenuated Interpretation

In this reading, which is taken up by the academic DPD without censoring it, siempre allows for paraphrases with roughly, more or less, so so, and other similar expressions that suggest an attenuated or hesitant assertion:

(12)

—¿Pasó muy mala noche?—Siempre; se estaba quejando algo al principio, pero le dieron unas pastillas y se durmió» (F. Calvo, Una historia armada, DPD, s/v siempre). ‘Did s/he have a bad night? Well, so so. S/he was complaining a little bit at first, but they gave him/her some pills and s/he fell asleep’

The adverb siempre is used here as an independent illocutionary form. The DA (s/v siempre) limits this usage to Colombia and Bolivia. Notice that this is a somewhat strange distribution, since these are not bordering countries, and those separating them (Ecuador and Peru) are, apparently, not included.

I do not have information about the origin of siempre in this interpretation. We cannot rule out the possibility that it is a calque of an indigenous language. In any case, there is another possibility: the passage from the emphatic to the attenuated affirmation that siempre shows in these cases might be somehow similar to the meaning change observed in the passage from seguro ‘certain, sure’ (also ‘of course, without question’) to seguramente ‘most likely’ in Spanish.

The semantic slippage from certainty to probability I am referring to does not occur in English (surely does not mean ‘probably’), but it took place early in Spanish, and is also registered in French (sûrement), in sharp contrast to Italian (sicuramente ‘definitely, surely’; not ‘most probably’). If we bear in mind that universal quantifiers sometimes form idioms of emphatic affirmation or intensification (de todas todas ‘all in all, without a doubt’; a toda costa ‘at any cost’, con todas las de la ley ‘fully-fledged’, etc.), the loss of emphasis or firmness that siempre presents in (12) could be seen as a particular case of a more general way to attenuate the forcefulness or the certainty of a statement. In any case, more work is required to confirm or disprove an explanation along these lines.

6. Conclusions

The main hypothesis defended here is the idea that most of the uses of the adverb siempre that we may call dialectal (in the sense of “not shared by all Spanish varieties”) maintain the original sense of this adverb as a universal quantifier, but differ in the quantified entity. The are good reasons to suppose that the continuative and the progressive-comparative readings, attested in RP Spanish, are direct calques of Italian, even if these uses are shared by French. On the contrary, the concessive-adversative interpretation seems to be attested in almost all areas of American Spanish.

I am well aware that the existence of similar uses of certain expressions of the same language in distant places may be interpreted as cases of polygenesis, rather than as the result of substratum, adstratum or other forms of direct external influence. In any case, this is an empirical issue. My research has focused on the semantic components shared and not shared by the different uses of a given adverb, which leaves for further work the important issue of explaining their history and their geographical distribution.

The durative and the event interpretations of siempre are shared by all Spanish speakers, as well as probably universal in languages with some equivalent of always. The non-selective binding interpretation of this adverb in generic sentences is also widespread (remember the example about planets and satellites), although the series created by siempre in the progressive interpretation may be non-temporal for some speakers only.

I have also argued that the concessive-adversative interpretation of siempre has important points of contact with the (non-dialectal) interpretation that I have called last resource reading. This reading (extensively studied in the French literature) is common to most Romance languages, as well as English. I have also argued that the fact that the latter interpretation of siempre is more syntactically constraint does not necessarily distance it from the concessive-adversative value.

The last of the readings examined (here called attenuated) is the least known and studied of all. I have tentatively suggested that it could be related to the attenuation registered in the use of certain adverbs of affirmation, but this is a line of explanation to be confirmed.

It is often said that the best way to understand the meaning of function words is to break them down into their constituent elements. In the case of siempre, this well-known strategy provides good results. In fact, the constituent elements relevant here are only two: a universal quantifier and a variable. In most cases, the latter corresponds to time points or events. But, as we have seen, there are cases in which the variable may be represented otherwise: by stages of some succession, by individuals that appear aligned in some progression, or even by the totality of the discarded options, whether or not expressly mentioned in the previous discourse. This makes the universal quantifier the fixed item in all interpretations (with the possible exception of the last one, pending further information), whereas the quantified element varies according to different uses and contextual adaptations.

The loss of the temporal value of siempre (and its equivalents in other Romance languages) in some of its uses might be further studied in several directions. For example, it would be interesting to find out whether the unselective quantification mechanism that allows us to quantify over a linear set of individuals forming series or sequences has a counterpart in grammaticalization processes. The concept of ‘succession’ or ‘series’ seems to remain, once it has lost its temporal value, in a relatively similar way as the prospective meaning remains when a deontic modal of obligation (as Lat. auere ‘have to, must’) becomes a future tense, a standard process in Romance historical syntax. The evolution of a grammatical word into another grammatical word constitutes one of the possible forms of grammaticalization. It is, then, therefore conceivable that the semantic retention or persistence that characterizes many processes of grammaticalization (Hopper and Traugott 2003; Verveckken 2015, among others) is also present in the passage from certain basic or elementary meanings of some particles to more abstract senses. From this perspective, it would not be surprising that always and its Romance counterparts cease to refer to a time series in certain contexts without necessarily losing their quantifying import.

The proposals introduced in this article open several interesting questions, most of which have been only schematically outlined. It goes without saying that I would be satisfied with the sole fact that the lines of research I am referring to are traced in the right direction.