The Spanish imperfecto may be used to update time in a narrative or a report. We take this phenomenon as a core property in order to account for a set of uses of the imperfective past tense-aspect form. We delimit these uses and sub-classify them according to their roles in the more abstract temporal structure. In order to do so, we apply the prominence-based account of temporal discourse structure by Becker & Egetenmeyer (2018). The findings of our analysis of the Spanish imperfecto are tripartite. First, we show that the typical narrative imperfect is only one instantiation of the updating imperfect and that there are also other interesting uses. Second, not all uses classified as cases of the narrative imperfecto in the research literature fall into our class of updating uses. Third, although there are unquestionable parallels with respect to the French updating imparfait, the assumption that the two forms have the same usage potential needs to be dismissed. We show that, within our more fine-grained categories, there are important differences between the two languages
The Spanish imperfective past tense-aspect form, the pretérito imperfecto, is generally assumed to be marked for imperfective aspect (see ). This property distinguishes the imperfecto (cantaba) from the pretérito perfecto simple (canté) (see , fn. 14 for references). The former indicates that only an inner part of the eventuality expressed by the verb in question is viewed, while the latter asserts the eventuality in its entirety (see ). However, as has been shown repeatedly, the imperfecto sometimes takes the place of the pretérito perfecto, as in the well-known example (1) from .
‘[1] At dawn the army set out, [2] crossed the mountain, [3] and soon after established contact with the enemy.’
In line with the French linguistic tradition (see , , , etc.), such uses are called “imperfecto narrativo” (see , ). Crucially, in clause [3], a telic verb (establecer, ‘establish’) follows a sentence adverbial expressing temporal posteriority (poco después, ‘soon after’) (see and for French; and and -q for Spanish). These three often-cited properties ― the expression of posteriority, the telicity of the verb in question and the co-occurring adverbial expression ― are central to what we call the “typical narrative imperfect” and translate into a basic but viable (though disputable, see fn. 3) definition of this specific use as we determine it. However, in the research literature, other structures are also discussed under the term of narrative imperfect, which are diverse in nature and do not always share these traits completely. The decisive characteristic for the classification generally put forward is the possible substitution of the imperfect for the perfective past (see , and others). For instance, examples like (2) and (3) are mentioned. In both examples, the verbs marked for imperfective past co‑occur with temporal adverbial expressions (en esos momentos, ‘at that time’; ayer, ‘yesterday’), which, however, do not express a temporal gap.
‘[1] On September 22, the meeting began in Vienna, the capital of Austria. [2] At that time, Iraqi President Saddam Hussein was launching a furious air and ground attack against Iran. [3] Thus began a war that would last eight years and cause tremendous material damage and great loss of life in both countries.’
‘The President of the Eurogroup, J.C. Junker, said yesterday that the effects that the tragedy of unemployment would have on the Spanish economy had not been properly gauged when calling for austerity.’
Example (1) seems to represent a case of the narrative imperfect upon which most scholars in the field would agree. We will call it the typical narrative imperfect. By contrast, there may be less agreement on the uses in (2) and (3) among scholars in the field. In our terminology, neither (2) nor (3) is classified as narrative imperfect. However, we classify (3) as a kind of updating imperfect, the classification we are most interested in (see below). As a translation into Spanish, we suggest the term imperfecto actualizador. The fact that, depending on the linguistic point of view, examples (2) and (3) may be classified as instances of narrative imperfects, makes it worthwhile to discuss their characteristics and delimit them within our classification system.
The main factor we will operate upon is the property of updating narrative time. and others similarly utilize the concept of narrative progression. However, we use the slightly more narrow and formally precise principle of updating the reference time, in line with . This property allows us to delimit certain uses of the Spanish imperfecto from others (see for a similar analysis of French data). We mainly focus on those uses which we classify as updating, and discuss their properties in more depth. However, as the delimitation along these lines seems to be new with respect to the Spanish imperfecto, we first present the account as it is the basis of our classification system.
Our view is based on the account of temporal discourse structure by . We employ this account with respect to the phenomena in question in order to deepen our understanding of the structural role the tense-aspect uses play in a discourse.
An important aim of this paper is to contribute to the understanding of an important set of potentially problematic uses of the Spanish imperfecto. As said, we discuss examples from the literature and classify them with respect to our categorization system. It seems that the Spanish narrative imperfect has received less attention than its French counterpart. There are, however, at least three extensive corpus studies on the Spanish case, namely by , , and . In addition, we apply our account to important sub-cases retrieved as part of our own corpus investigations using CREA. Overall, we focus on temporal structuring and not on connotations or stylistic properties of certain uses of the imperfecto. We mainly focus on uses pertaining to the foreground and exclude cases of co-temporality and habituality, and also cases where the imperfecto occurs in embedded sentences (see the argumentation in Section 3).
We analyze data retrieved from the CREA corpus. Our corpus queries are based on a large set of verbs with properties susceptible for updating the narrative time. The set mostly consists of action verbs expressing, for instance, movements, speech events or political acts. Importantly, we exclude cases where the imperfecto is used for grammatical reasons (for instance, in embedded speech) or other rigid relational reasons (like co-temporality) (see Section 4 for details).
The paper is structured as follows. In Section 2, we outline the account of temporal discourse structure by . We begin with this because our classifications and analysis are based on it. In Section 3, we discuss the nature of a variety of cases from the research literature and determine their position with regard to our categorization system. Section 4 is dedicated to data from our own corpus study. Finally, in Section 5, we summarize our insights.
In , we present an account of temporal discourse structure which is able to cover many properties of tense-aspect use in texts. The account has been developed with narrative texts in mind, as can be noted in some parts of the terminology, but it can also be applied to other genres. It comprises three different levels, namely: (i) linear relationships between times, (ii) perspectivization phenomena, and (iii) the foreground and background distinction (see ). Crucially, on all three levels, the account covers the horizontal plane, i.e., linearity, and also the hierarchical relationships of the abstract structure corresponding to the temporal constellations analyzed (see ). The operationalization of terminology and structures is a development of discourse representation theory (DRT; see , Kamp & Reyle 1995, , etc.), which gathers some inspiration from the work of . We have shown how to apply the account in several papers (see , , etc.). Among those papers, concerns the updating use of the French imparfait. The main aim of the present paper is not to contrast French and Spanish, although we mention several interesting discrepancies (see for a detailed comparison of narrative tense use in the two languages). Rather, we seek to draw a partially new picture of the Spanish updating imperfecto, grounded in the discourse view. We thereby determine its discourse potential and distinguish fine-grained divergences between specific uses. In the following, we will briefly introduce the terminology used by . We will elaborate on the crucial property of updating the reference time. Finally, we will introduce the notion of prominence (see ).
We account for temporal structure in texts by taking the different temporally relevant components of a text and mapping them onto an abstract structure (see for further details of the following outline). The linguistic components and indicators of temporality are manifold, ranging from different sorts of more or less explicit lexical material such as adverbial expressions and actional classes of verbs (Aktionsart) to TAM forms. In order to build the abstract structure (see the visualization below in Figure 1), we concentrate on five different types of time points and time spans. Some of them have been a central part of the research on temporality for a very long time (see ), but some have undergone important precisions since then (see and later publications in DRT), to which we have contributed further refinements (see , ). The speech time (n) determines the origo of the speaker and thereby sets the basis for all temporal relationships. The time of the eventuality (e) is the potential extension of the eventuality (see ). The location time (loc), by contrast, is the time span actually allocated to an eventuality according to the text the corresponding expression occurs in (see ). Adverbial expressions may render the location time explicit (see ). The reference time (R) has already been introduced by , but is re‑defined as a discursive concept in DRT. It is an abstract conception of the time which has been arrived at when an eventuality is asserted in a text (see, for a similar formulation, , with reference to , and Kamp 2013: 119). Formally, the reference time is introduced via a corresponding location time (see ). Crucially, in the account, only eventive but not stative eventualities temporally develop the narration further, specifically because they lead to the introduction of new reference times (see with reference to ). We adhere to the so-called “pulling account” (see ), which means that in the structure, the location time, corresponding to a sequentially following eventuality, is anchored to the previously introduced reference time, and consequently introduces its own reference time (see and the visualization below). This shifting of the reference time is what we call updating. The fifth time point completing the relevant set of the account is the perspective time (PT). It is the time point from which an eventuality is viewed (see Kamp 2013: 119). In the case of typical narrative sequences it is “anchored to the speech time” (Egetenmeyer accepted b, with reference to ). In order to account for the cases of interest in the present paper, we will mainly deal with the location time and the reference time. Before detailing our analysis of the updating imperfect, we first discuss a short extract involving the pretérito perfecto simple (see (4)), in order to show how the concepts above are applied in the case of temporal updating.
‘[1] On my birthday my father went down to the bakery on the corner [2] and bought the finest cake he could find.’ (, chapter 7)
In line with the visualization technique first presented in , the example can be represented as in Figure 1.
Example (4) consists of two main clauses with perfective past tense-aspect forms (bajó, ‘went down’; and compró, ‘bought’). For the sake of simplicity, we ignore the third verb form (encontró, ‘found’). Despite its perfective form, it does not develop the narrative time further, which has to do with its syntactic realization as part of a relative clause. As the eventualities expressed are asserted for some time in the past, Figure 1 shows the respective structure to the left of the speech time, which is marked as n (see and for this and the following details). The first clause is introduced by a temporal adverbial (el día de mi cumpleaños, ‘on my birthday’) which sets the time frame of the two events. In Figure 1, this is indicated by the dotted brackets, which represent the location time of the adverbial expression (locAdv ). The location time of the first event (loc(e1)) is temporally anchored to the location time of the adverbial, as the double-headed arrow indicates. The verb phrase bajó (al horno) (‘went down (to the bakery))’ expresses a bounded event viewed in its entirety (with [+telic] Aktionsart and [+perfective] aspect). Thus, it is eventive in nature, and therefore, via the corresponding location time (loc(e1)), a reference time (R1 ) is introduced into the structure. In terms of temporal structure, the following event (compró, ‘bought’) depends on the first event. Formally, its location time (loc(e2)) is anchored to the reference time (R1 ), introduced as a consequence of the eventive character of the first eventuality. In Figure 1, this is again shown by a double-headed arrow. The second eventuality is also eventive and via its location time a new reference time is introduced (R2 ). The new reference time R2 , in turn, would be available as an anchor for subsequent events.
In , a typical use of the imperfect, for instance in the case of a description, is analyzed as a state. In the temporal structure, the corresponding time span is said to include a previously introduced reference time (see , partly quoting ). presents the fragment in (5) to show what he intends as an anaphoric use of the imperfecto. As he points out, the situation expressed by the verb marked for the imperfecto in sentence [3] can either hold true at the time of the event referred to in [2], or at the time of both events in [1] and [2] together ().
‘Juan separated from her. A few months later he quit his job. He was not happy.’
In Figure 2, we represent the interpretation according to which the state expressed in [3] holds true at the time of the second event (from sentence [2]).
With respect to sentences [1] and [2], Figure 2 is similar to Figure 1. The two sentences refer to events which update the narrative time. In the visualization of the temporal structure, this is again represented by the introduction of two reference times (R1 and R2 ). The verb marked for imperfecto in sentence [3] does not update the narrative time. Therefore, there is no new reference time, which would be introduced via the corresponding location time (loc(s3)). Rather, loc(s3) takes R2 as its antecedent, as visualized by the dashed line circle.
The above example shows the opposition between the pretérito perfecto simple and the imperfecto already mentioned in . Only the perfective updates the narrative time, that is, shifts R to the right; the imperfective form does not (see ). The basic problem of the narrative imperfect, in a broad sense of the term (see the next sections for terminological refinements), is that it deviates from this standard tendency and does, in fact, update time. This will be the basis of our analysis here. What is more, it allows us to delimit important uses of the imperfective past tense-aspect form. In Section 3, we will recategorize some cases from the literature.
As temporal adverbial expressions are an important part of our analysis, several properties and distinctions need to be considered in order to determine their role in the discourse structure. First, functionally, they may either localize the narrative time or they may render a time point (more) specific. This distinction is brought up in Becker & Egetenmeyer ( with reference to and ). They present the example pair (6) and (7).
‘The car crash happened at ten o’clock. (Adapted from )
‘Yesterday Fred bought a lawn mower.’ ( with reference to )
In (6), a las diez (‘at ten o’clock’) contributes temporal specificity to the location time of the event (see ). By contrast, a sentence adverbial (see, for instance, ) like ayer (‘yesterday’) in (7) introduces its own location time within which the buying event takes place (see ). discusses the second case as “adverbios de marco” (‘framing adverbs’) and points out that the class is determined by the syntactic position and the semantics of the adverbs in question. For instance, mañana (‘tomorrow’) in (8) is not a framing adverb despite its parallel content with respect to ayer (as in (7)) (see ). Among other tests, she presents the negation test, in which the scope of negation is decisive (see ). In (8), a negation is already present. Importantly, no (‘not’) has scope over the adverb (see ). In our terms, only ayer (‘yesterday’) in (7) introduces a location time, while a las diez (‘at ten o’clock’, in (6)) and mañana (‘tomorrow’, in (8)) only specify an already introduced time point further.
Second, an important factor are the semantic and referential properties of temporal adverbial expressions. The following three classes are especially relevant within this study and they fulfill very different roles in discourse. A first class of adverbial expressions refers to a kind of absolute time in the real world or the non-linguistic discourse world, for instance, en 2023 (‘in 2023’). A second class refers to a time co-dependent on the discourse world and cannot be interpreted independently of the speaker’s origo (see ). Such adverbs are often called deictic (see ). Typical examples are now, yesterday, in two weeks (see ). Furthermore, also at ten o’clock or on Sunday are context-dependent (see ). The adverbial expressions of a third group only relate time points and time spans within the discourse. Examples are then or next, which may have the mere function of indicating sequentiality (see for several nice examples). However, also the already mentioned adverbial expressions which introduce a temporal gap pertain to this class (for instance, poco más tarde, ‘a little later’).
In addition to the “vertical”, that is sequential relationships between times, the account considers hierarchical structuring. This is covered under the term of prominence. Prominence is understood, in the sense of and , as the property of an entity “sticking out” from a set of equal entities (see also ). What is especially important for temporal discourse structure is that prominent entities have the potential to “serve as anchors” () within their domain. In narrative texts, the domain basically equals an episode (see ), but the concept can be covered more precisely in terms of right frontier constraint (see for details). The prominence value is measured quantitatively in terms of the amount of anchoring relations a time point licenses, and the domain restriction allows a certain time point to be determined as maximally prominent (see ). Time points which take part in regular chains of anchoring, that is, which occur within a sequence, have a prominence value that is equal by default (see ). If a reference time is not available as anchor, its prominence value is said to be minimal (see ).
The final ingredients necessary for our account of discourse structure are the semantic and pragmatic relationships between propositions. In this respect, we mainly use the terminology and conception of , who term these kinds of relationships rhetorical relations. For instance, when events in sequence are expressed, the respective propositions are said to be connected by the narration relation, while a proposition explaining a preceding one holds an explanation relation with respect to the preceding proposition (see for details). Importantly, rhetorical relations may be either coordinating or subordinating (see ; ). The narration relation is a coordinating relation; explanation and elaboration, by contrast, are subordinating relations (). As subordinating relations do not attach information at the main level of a story line, they cannot push the main events forward temporally.
The Spanish imperfective past tense-aspect form imperfecto has been investigated in various and extensive ways (see , ). For instance, it has been approached as part of the tense-aspect paradigm in general (see , ) and has been described in terms of vectors (see ). In addition, it is often addressed in structurally or pragmatically embedded contexts. In such contexts, sequence of tense phenomena are a central factor of influence (see ). A related phenomenon of high interest in temporal semantics is double access, that is, the potential of complement clauses to take the time of an introducing speech verb marked for past tense or the moment of speech as antecedent (see , ). Some researchers also focus on the potential uses of the imperfect beyond the expression of temporal relationships (see , ). Other interesting studies concentrate on the combinatory possibilities of the imperfecto with Aktionsart properties, adverbial expressions and verbal periphrases (see Dessì Schmid , , , ). Finally, another line of research treats the use of the imperfect in context. In Romance linguistics, an influential textual analysis of the imperfective past is , who relates the imperfect to the background. Among the studies in this realm, some focus on a more specific genre, such as newspaper texts (see , Azpiazu Torres , ).
Crucially, we are interested in a set of uses of the imperfective past tense-aspect form, some of which are covered under the term imperfecto narrativo. Some of the above-mentioned publications and others mention the narrative imperfect. However, there are only a few publications which analyze the Spanish narrative imperfect more thoroughly. and present extensive corpus studies on the matter. Some of their data and the data of others will be discussed below.
An important part of the research literature on the Spanish narrative imperfect applies general insights coming from research on French, for instance, by , , and (see for instance, , , ). The list of properties of the narrative imperfect proposed by is especially thorough and generally accepted. It includes the following points: the narrative imperfect may be substituted by a perfective past as it shares the property of the perfective past of advancing the narrative temporally; often, it is preceded by a temporal adverbial; and it tends to mark a telic verb (see also Bres , and many others). As we will see in the present section, some of the examples cited in the research literature do not meet these criteria. However, our own analysis sets out from a different angle by taking temporal advancement as a basic ingredient for the uses investigated. The most important competing readings often require a wider context in order to be determined, namely, co-temporal and habitual uses.
In Section 2, we presented our account of temporal discourse structure, which allows us to distinguish between updating and non-updating uses of TAM forms. This will be applied in the following. In our view, the narrative imperfect is only one instance of such uses. Crucially, however, the updating property is necessary in order for an imperfect to fall into the category of narrative imperfects. In the following sub-sections, we will discuss a set of examples classified in the literature as cases of the narrative imperfect, or with competing terms like perfective imperfect (see ) or punctual imperfect (see ). First, we account for quite diverse cases, among which many do update narrative time (§ 3.1). In § 3.2, we focus on examples which do not update time or where the imperfect is a grammatical reflex. In § 3.3, we discuss the distinction with respect to progressivity and show that in this realm there is the potential for ambiguity.
We mentioned the case of the typical narrative imperfect in Section 1, where we presented an example from . Further examples of this kind are (9) and (10). The central properties are that the verb marked for the imperfective past tense-aspect form is embedded under a preceding sentence adverbial expressing a temporal gap. In this case, the updating of the reference time crucially depends on the location time of the adverb (see § 4.1 for details and a visualization). In these two examples, the verbs (or rather, the verb phrase in (9)) are telic in nature. However, a short temporal extension (achievement verb) does not seem to be necessary, as we only find this in (10).
‘Shortly after, he was shamefully driven to prison.’
‘Shortly afterwards, the bomb exploded.’
In fact, and present cases of the narrative imperfect to argue against the aspectual hypothesis (see for a way of capturing such cases with the formalism of vector analysis). They suggest that their example (10) is licensed by an interpretation which actually involves the embedding of the imperfect under another verb, as in (11). However, the pervasiveness of the typical narrative imperfect and also of other updating uses of the imperfect clearly counter this intuition.
‘A short time later, they watched in horror as the bomb exploded.’
The literature mentions highly interesting examples involving speech verbs. Classifying them as to whether they update time or not is not always easy. Among the cases which update time, we distinguish between a group in which there is an update on the main level of the story line and a group in which the update pertains to a rhetorically subordinate structure (see below). In order to distinguish occurrences along the lines of the two groups, a certain amount of preceding context is necessary.
In example (12), originally from , but presented in the adaption by , the update is realized as part of the main structure. The main clauses in [1] and [2] express a sequence of two speech events. Thus, there is a temporal update which can be formally captured as in Section 2. This is a very nice case of the often-mentioned property that such imperfect forms can be substituted by a perfective past tense-aspect form.
‘[1] The observatory announced that a hurricane was approaching our coasts in a NE to SW direction. [2] The weather report added that the first gusts would reach the island early this morning.’
Example (13), by contrast, shows an update as part of a rhetorically subordinate structure. In order to show this, we included sentence [1], which precedes the clauses involving verbs marked for imperfect. Importantly, clauses [2] and [3] express an elaboration of what is said in [1]. A similar principle is mentioned in under the term “macroevento” (‘macro event’), where co‑temporality is given with respect to a more global level, which, however, may only be implied (see also ; furthermore, use the term in the context of habituality). In the case of (13), this leads to two crucial properties. First, structurally, the location times of the two subordinate events in [2] and [3] are temporally parallel to the time span of [1]. More specifically, within the subordinate structure, [2] does not update time, but [3] does. However, given this co-temporality resulting from the property of the elaboration of sentence [1], we cannot argue for the imperfecto as a mere choice. Rather, it may be said to be employed as a grammatical consequence. Therefore, we do not consider such cases in our analysis in Section 4.
‘[1] In an apparently Solomonic decision, the Government put an end yesterday to Juan Ortiz Úrculo's brief but strident eight-month term as General State Attorney. [2] Shortly before the meeting of the Council of Ministers, President José María Aznar communicated his decision to the Minister of Justice [3] and announced the name of his future successor.’
However, it should be noted that this constellation seems to be relatively frequent. It is also the case in (14). The example underlines how vast the classification of speech verbs may be understood to be. The verb desempolvar (‘dust off, revive’) does not normally designate a speech event either in its concrete or in its figurative sense. But in its use in clause [3], it makes reference to an event which is instantiated in the non-linguistic reality as speech. Again, the sentence preceding the part cited in the research literature (see ) is necessary to determine the status of [3] and [4]. They elaborate on the preceding context. In the given context, the adverbial expression ayer mismo (‘just yesterday’) underlines the elaboration relation. It expresses that [3] and [4] are to be understood as exemplifying the previous content.
‘[1] As expected, the text raised blisters among feminists, homosexuals and progressive sectors of the Basque judiciary and politics. [2] They believed that their reflections “distilled conservatism” and that they were a clear sample of “inadmissible” fundamentalism in a modern society like the Spanish one. [3] Just yesterday, the association Judges for Democracy revived the text [4] and used it as an argument to show their skepticism.’
The above examples potentially involve a sequence of events. However, speech verbs (and other verbs ― see below) may also occur individually as punctual and possibly updating cases. The speech verb in (15) is clearly telic (see , with reference to ). It may be expected to have its own reference time associated with it. However, in order to determine whether it updates time or not, the context needs to be accounted for. Depending on the context, a possible alternative interpretation would be the instantiation of an imperfect introducing a “special event” (or a “date setting imperfect”; see ; see also below and § 4.2).
‘The President of the Eurogroup, J.C. Junker, said yesterday that the effects that the unemployment tragedy would have on the Spanish economy had not been properly gauged when calling for austerity.’
Another relevant case presented in the research literature is found in example (16). presents it to exemplify the “pretérito biográfico” (‘biographical preterit’). This classification concerns the topic, and there is potentially even a genre where this use occurs. However, the content per se does not necessarily (or even not normally) determine the temporal structural functioning of a form and thereby the corresponding systematization. However, the example presented is an interesting one, and a more detailed analysis yields relevant insights. It consists of two parts. In both parts the reference time is updated. Structurally speaking, clause [1] is another instance of a typical narrative imperfect (see above). Clause [2] is superficially similar, but the adverbial expression (el 8 de septiembre, ‘on September 8’) does not designate a temporal distance, although it may imply one. We relate this instance to what calls “imparfait d’événement” (‘imperfect of occasion’). Although this is possibly not the standard case, in (16) it is part of a sequential relationship. We come back to this in § 4.2.
‘[1] A month later he was obliged to stay in bed [2] and on September 8 he died at his home in Garmisch.’
Next, we focus on cases which neither depend on an adverbial expression nor involve a speech verb. In (17), the verb marked for imperfecto in [3] (ponían, ‘seized’) contributes a temporal update. It is a general action verb. The only possible indicator of sequence is the conjunction y (‘and’). It should be mentioned that the verb in [2] (cercaban, ‘surrounded’) does not necessarily correlate with a temporal update. In this clause, the adverbial expression en el mismo instante (‘at the same moment’) lexically marks co-temporality. However, a sequential or a developing reading also seem to be available. There are two reasons for such alternatives. First, pragmatic inferences may lead to a sequential interpretation. This involves reinterpreting en el mismo instante as allowing the expression of an adjacent sequential relationship. And second, the sentence may be taken to convey textual development (in the sense of Egetenmeyer accepted a). In terms of the scale of textual development presented in Egetenmeyer (accepted a), it may be classified as a case of “furnishing update” that borders on being an “argumentative update”. The simplified reasoning would be that the situation is enriched by [2], which also introduces a contrast with respect to [1].
‘[1] He wanted to take up his weapons, [2] but at the same moment, obedient to a signal, the mastiffs of the guard surrounded him [3] and they seized him.’
A final case we want to mention here is the time updating use of the imperfecto in football reports. This is mentioned, for instance, in . A more thorough coverage, with a significant amount of data, is presented in . However, she relates it more generally to the sports sections of newspapers (see ). According to a previous study on French, the football context brings about the most extensive substitution by the imparfait of an alternative perfective past tense-aspect form (see , with reference to Labeau ; and others). As (18) shows, the Spanish imperfecto may also easily mark achievement verbs in sports reports. A temporal update by means of marcaba (‘scored’) is probable; however, in order to determine the status of such a clause without ambiguities more preceding context would be necessary. The verb phrase in [2] (arrancaba gritos […], ‘provoked shouts […]’) introduces the consequence of the goal in [1]. However, the subject referent does not change between [1] and [2]. Furthermore, a basic understanding of the situation would be that it is the goal which has the audience react. Therefore, strictly speaking, the subject referent realizes one only action. The reaction itself is not introduced verbally (gritos y aplausos, ‘shouts and applause’). However, the basic abstract temporal structure concerns the verbally introduced eventualities. Consequently, in this strict view, the second event (arrancaba, ‘provoked’) is captured as co-temporal. Therefore, in contrast to , we do not take it to be an instance of an updating (or narrative, in their terms) imperfect.
) presents many examples from sports reports. She thereby underlines that there may be a special tendency to use the imperfecto in this text type. However, it seems that many of the examples can be explained as (less obvious) cases of co‑temporality or as elaborations in the sense discussed above. In (19), the first imperfect form alcanzaba (‘achieved’) does not have to be but may be a consequence of the elaboration relation that [2] holds with respect to [1]. Similarly, [3] elaborates [2], and [4] elaborates [3]. In terms of temporal relationships and relative temporal extension, the verbs in the respective following clauses always express a sub-part of the sentences preceding them. That is, the eventuality conquistaba (la “Champions League”) (‘won (the Champions League)’) in [3] is only a part of what is expressed in [2] about reaching the biggest success. Similarly, in [4], winning the final match is only a part of winning the Champions League ([3]). Thus, [3] and [4] hold a temporal inclusion relationship with respect to the eventualities preceding them. This grammatically licenses the imperfecto.
‘[1] In Germany, there is a prize for the one who solves the paradox: Borussia Dortmund, the reigning European champions, are currently in the relegation zone; Kaiserslautern, a newly ascended team, are the sole leaders. [2] Last spring, Borussia Dortmund achieved its greatest sporting success. [3] After winning two league titles in a row, they won the Champions League. [4] They beat Juventus, a previously unbeatable side, 3-1 in the final.’
In the preceding section, we saw that the research literature discusses many different uses and cases of the imperfect which our account also explains as cases of the updating imperfect. We also saw several cases which we would rather not classify as updating. The factors leading to ambiguity, then, are subordinating rhetorical relations, specifically the elaboration relation and less obvious cases of co‑temporality. In the current section, we discuss further cases presented in the literature as narrative and perfective uses, but which do not update the reference time in the sense presented in Section 2. The competitors in this case are again co-temporal relationships.
We begin with a simple case where the original author is apparently interested in a different property than the one we are focusing on. includes (20) in the group of “imperfectos perfectivos” (‘perfective imperfects’). The property he focuses on is that the right boundary of an eventuality (represented in [1]) is made explicit by means of a following adverbial clause introduced by antes (‘before’) (as in [2]) or hasta (‘until’) (). However, with respect to the temporal updating potential of the verb marked for imperfective past, such a following adverbial clause is not crucial. In addition, in examples like (20), [1] cannot contribute a temporal update to the temporal structure.
‘[1] I was happy [2] before you arrived.’
Example (21), from , contains two telic verbs marked for imperfecto. Such cases are quite frequently classified as narrative imperfect in the literature (in line with standard characterizations; see, e.g., ). However, neither [2] nor [3] yields an update of the reference time. In fact, in both sentences, an adverbial expression indicates their co‑temporality (en esos momentos, ‘in those moments’; and así, ‘in this way’). They also license the occurrence of the imperfecto grammatically.
‘[1] On September 22, the meeting began in Vienna, the capital of Austria. [2] At that time, Iraqi President Saddam Hussein was launching a furious air and ground attack against Iran. [3] Thus began a war that would last eight years and cause tremendous material damage and great loss of human life in both countries.’
Finally, example (22) presents a property description. The verb marked for imperfective past (era, ‘was’) does not introduce an event via which a reference time could be introduced into the temporal structure. mentions the possibility of classifying it as a case of the narrative imperfect. This might be due to the explicitly marked temporal delimitation. We do not share this intuition. More importantly for our discussion, the verb in the clause cannot be involved in a temporal update. The referential expression aquel (‘that’) clearly indicates that the location time corresponding to era (‘was’) includes a previously introduced reference time. It is a state in the sense of .
Habituality is one of the core meanings of the imperfecto (see ). Interestingly, within habitual sequences, eventualities may be described as taking place one after another. However, first, habitually occurring events are said to be linked to the background of a plot (see ). And second, the respective eventualities appear under the scope of a habitual operator. As it is this operator which yields the realization of an imperfective past tense-aspect form, the imperfective marker may be said to be a grammatical reflex of habituality. However, when working with corpus data, it is not always easy to determine whether a sequence of verbs marked for imperfective is habitual or not. The reason is that such discourse fragments may be very long and a corpus query may return strings too short to include the relevant habitual marker. This may have been the problem in example (23). presents clauses [2] to [5] as a “contraejemplo[…]” (‘counterexample’) to the standard distribution of a foregrounded and actualizing pretérito perfecto simple and a backgrounded, descriptive imperfecto. However, if we include the beginning of the paragraph concerned, seven sentences before the example extract, it becomes apparent that the journey described is a habitual undertaking. Therefore, the imperfective marker is the grammatically necessary form and the example is not relevant for our kind of analysis.
‘[1] The journey from Rome to Bomarzo, which we undertook to avoid the heat and the malignant fevers and which ceased in 1528, because then we settled permanently in the castle, was awaited with emotion by the writer of these pages. [2] We crossed the First Gate, which the ancients called Saxa Rubra, [3] and continued along the Via Flaminia, passing near the arch that Constantine erected to commemorate the place where he camped before the battle against Maxentius; [4] we skirted the rock of Cesare Borgia and the countryside overlooked by the summit of Mount Soratte, [5] and thus came to rest in Civitta Castellana, where we prayed in the cathedral, gleaming with mosaics.’
Finally, we want to discuss a few more examples from the literature which are ambiguous in terms of temporal updating. The competing property we focus on is progressivity. Interestingly, such cases are sometimes presented by means of a term which is ambiguous in itself. While the term “perfective imperfect” (see, for instance, ) is transparent and can be supported by a substitution test with the perfective past, the term “imperfecto puntual” (, ‘punctual imperfect’) does not determine how the punctuality is realized. Not only can an eventive (and updating) eventuality be punctual, but also a progressive, as it also highlights a specific time point. More specifically, the progressive is said to be monofocalized (see ). We intend progressivity as a semantic category, not in terms of a formal marker (see for the distinction). The progressive reading is defined in as referring to “an event which is in progress at a particular point in time” (see also ). The crucial difference with respect to the TAM use we focus on is that progressives generally do not shift the narrative time but take a previously introduced reference time as their anchor (see with reference to , and ). Bertinetto (, ) distinguishes a further category similar to the progressive which he terms “continuous”. It parallels the progressive in its semelfactive nature and the expressed eventuality is asserted for an extended time span (see ). Thus, adverbial indicators of the progressive would be punctual (at 5 o’clock) and extended with the continuous (during the lecture) (see for the distinction and examples).
In the following, we discuss two examples from the literature which, in light of the reduced context given there, may have an updating or a non-updating, but rather, progressive, reading. There are no generalizable overlapping properties between the two groups. However, introduces another kind with features of both types called the “progressive with narrative features”. We will come back to it in the end of this sub‑section.
Example (24) is presented in as a case of punctual imperfect lacking an adverbial, which would otherwise express “sucesión rápida” (, ‘quick succession’). Due to the lack of context, (24) is ambiguous between an updating and a progressive reading. The sentence adverbial el día convenido (‘on the appointed day’) determines a time point, which in standard literary narratives may be expected to be posterior to what has been mentioned previously. Therefore, the updating reading is possible. However, in contrast with the adverbial expressions of a temporal gap, the updating reading is not necessary. In addition, the inserted temporal expression como a eso de las cuatro […] (‘at about four o’clock […]’) is punctual. This time point might be highlighted in the story, which would allow for a progressive reading (in the sense of a las cuatro los dos jesuitas estaban saliendo del Colegio, ‘at four o’clock, the two Jesuits were leaving the college’). We do not vouch for the progressive reading having a high probability, but we want to stress that without context it cannot be excluded here.
‘On the appointed day, at about four o’clock in the afternoon, the two Jesuits left the college, heading for the village.
A similar potential ambiguity arises in (25). The reasons are similar. The adverbially conveyed time a la hora que indicó Novaliches (‘at the time indicated by Novaliches’) may again be part of a temporal updating. However, the temporal extension expressed is sufficiently reduced to be compatible with a progressive reading. Note that the verb in the pretérito perfecto simple (indicó, ‘indicated’) expresses a pre‑past temporal reference (see for substitutions of the pluperfect with the perfective past in Spanish as compared to French).
‘At the time indicated by Novaliches, or rather a little earlier, Beramendi’s car stopped at the Puerta del Príncipe.’
There are also certain cases involving verbs of speech which are ambiguous with respect to progressivity. This seems to be the case in (26). Crucially, the imperfect does not update the time. Rather, the speaker intends to re-locate the conversation focus to a past (reference) time, which is part of the common ground shared by speaker and hearer (see also , who, however, does not discuss the progressive component). At that past point in time, the saying event was ongoing. The functioning is underlined in subordinate structures involving como (‘as’) as in como decíamos antes (‘as we said/were saying earlier’) (see ).
‘I was telling you … a while ago that … tonight will be a night to put large doses of coffee in a thermos flask’
discusses examples involving the French imparfait which combine the otherwise mutually excluding features of updating and progressivity. They are termed “progressive with narrative features” and distinguished into two sub-cases, of which the first marks the update explicitly with mais + déjà (‘but + already’) (see (27)), while the second does not involve a marking (see (28) below) (see ). Example (27) begins with a state that the ego narrator is experiencing ([1]) during a previous passage of represented speech (see for this analysis of the example). In [2], however, the ego narrator is in a different situation, namely, the ongoing action of the prison chaplain being taken out of his hands (on m’arrachait l’aumônier des mains, ‘the chaplain was being snatched from my hands’). The change of situation is made explicit by mais (‘but’), while the progressive component is underlined by déjà (‘already’). In terms of temporal structuring, there is an updated reference time corresponding to the verb marked for imperfective past in [2] (R2 ). This reference time is taken as an anchor by the location time corresponding to the following eventuality (loc(e3), menaçaient, ‘threatened’) (see for further details, including a visualization of the temporal structure).
‘[1] I was nearly suffocating screaming all this. [2] But the chaplain already got snatched from my hands [3] and the guards threatened me. [4] The chaplain however, calmed them down.’ (Translation taken from )
Example (28) shows the case of the “progressive with narrative features without explicit marking” (). Again, a temporal update is conveyed via the verb marked for the imperfective in [3] (appuyais, ‘leaned’) with respect to [2], where R3 follows R2 . Sequentiality is necessarily given, as clause [2] refers to the movement towards the wall, where the leaning event in [3] can take place (see ). The eventuality in [3] is intended to have a progressive reading, as it is an ongoing situation extending beyond the reference time R3 introduced into the temporal structure via the location time corresponding to appuyais in [3]. shows a crucial difference of the progressive with narrative features without explicit marking compared to the structure with explicit marking, namely that the former introduces a reference time which is not available as anchor (minimal prominence). This has the consequence that the following location time (loc(e4)), corresponding to ai entendu (‘heard’), is anchored to the preceding reference time R2 and not to R3 (see Figure 3 below) (see ). This captures the fact that the peripheral listening has already started during the movement of the ego‑narrator towards the wall (clause [2]) (see ).
‘[1] Now I turned away [2] and I walked over underneath the little window. [3] I leaned (was leaning) my shoulder against the wall. [4] Without really following him, I heard that he started questioning me again.’ (Translation taken from )
The temporal structure of (28) is visualized in Figure 3 from . The minimal prominence is indicated by the brackets around R3 . As the arrows show, R3 is not taken as an anchor by loc(e4). Therefore, the eventualities can overlap temporally (see for the analysis).
Recently, this case has been taken up in . In a similar vein to the classification as a “progressive with narrative features” from , he treats the use of the imperfect in clause [3] as showing a “propulsive continuous aspect” (). The “propulsive” property is defined as “the dynamic advancement of the plot” (). In defining the term he refers to , but he does not employ their (or our) formalism (see ). A “continuous” reading, as mentioned at the beginning of the sub-section, is similar to the progressive in that it is “semelfactive”, and differs from it in referring to a more extended time span (see ). Thus, this classification nicely captures the properties of the example as described in .
We conducted an extensive corpus study, considering crucial properties of the form in our corpus queries. The aim was to gather further insights into the different sub‑kinds of the updating imperfecto in natural language. According to the literature on the narrative imperfect in French (, , , and many others) and Spanish (, , , ; etc.) a very important, and almost necessary, factor for the occurrence of the imparfait narratif is the telicity of the verb(s) involved. This has also been confirmed by our own study on French (). Another highly relevant property, which, however, the research literature implies rather than states as such, concerns the status of the participant involved. It is typically human and, due to the verb types affected, is often represented as agent. On these grounds, we selected a set of telic action verbs for our queries. Among others, for a possibly rather plain conceptualization, we used verbs with denotations pertaining to the concrete non-linguistic reality, or, more specifically, movement verbs (see also for this principle). In order to highlight telicity, we used achievement verbs like entrar (‘enter’), salir (‘go out’), llegar (‘arrive’), etc. For some of these verbs we also added their combination with an adverbial expression like […] más tarde (‘[…] later’) or después de […] (‘after […]’) in order to specifically gather examples of the typical narrative imperfect.
shows that speech verbs are a relatively frequent verb class marked for the imperfective past in perfective contexts. Therefore, we also included several verbs of speech or thought, like comentar (‘comment’) and recomendar (‘recommend’). We did not include the general speech verb decir (‘say’) because of its overwhelming frequency both in and beyond perfective contexts (see for an insight on the distribution of verbal types). For instance, in literary data, the habitual uses, which are also very frequent, would have to be excluded by hand. Furthermore, we used queries of what we call political acts, which may but do not have to refer to speech events, like decidir (‘decide’) and inaugurar (‘inaugurate’). In addition, we included verbs designating events in a football match, for instance, marcar (un gol) (‘score (a goal)’) and centrar (‘to center’). The verbs of speech, of political acts and coming from the football context are also telic. Finally, we also included a set of verbs of movement, whose basic meaning pertains to the Aktionsart of activity (atelic). However, this part of the investigation hardly returned any relevant examples.
We used the CREA corpus for the data search. We focused on data from Spain, but did not exclude other varieties. We collected relevant cases from among the first fifty hits retrieved by means of our queries. We sub‑categorized certain queries for literary or newspaper data. Beyond that, we did not restrict the queries. Most of the properties relevant in our analysis cannot be pre-determined in standard corpora (see below). Rather, the results of the queries were filtered manually with respect to the characteristic of the subject referent as being human, as well as several further properties which may hinder an updating reading or at least reduce its probability. In this process, other frequent uses of the imperfecto were ruled out. First, as mentioned earlier, we are not interested in cases of co-temporality and habituality and other world-related shifts. In these cases, the imperfect occurs for specific reasons which we do not address in this paper. Second, the verb form retrieved should occur in a main clause. This excludes the possibility that the inflection is the consequence of a grammatical mechanism, as in the case of sequence of tense phenomena (see ). In addition, an occurrence in a relative clause may have a property reading or other stative features (see ) which would also oppose an updating reading. Third, the eventuality expressed should pertain to the foreground. Otherwise, the grammatical form may simply be a mark of background (see ). Fourth, wherever possible, we retrieved cases of several imperfective forms in a row. This may underline the realization of an update by the imperfective form itself. The lack of restrictions in our query allowed us to gather a wide variety of cases and contexts.
The corpus study yields several highly interesting updating uses. In the following, their different properties are discussed. A special focus is placed on their updating potential. Furthermore, the Spanish imperfecto uses are compared with their potential French counterparts when it is necessary to motivate a certain sub‑type or when the comparison is helpful in order to describe the Spanish instantiation of one of the sub‑types more fully. We begin with two cases discussed in the research literature, the typical narrative imperfect (§ 4.1) and the imperfect introducing “special events” (§ 4.2). The two share the important structural parallel of being embedded under a sentence adverbial. The other three cases are not embedded. In § 4.3, cases from literary data are considered, while § 4.4 focuses on newspaper data and, more specifically, on the case of football match reports. Finally, in § 4.5, we discuss speech verbs marked for the imperfective past and their updating potential.
In line with , we define the typical narrative imperfect as a use which is preceded by an adverbial expression acting as a sentence adverbial under which the verb is structurally embedded. The adverb temporally relates the relevant clause to the preceding context by means of a specific time span. More specifically, it introduces a “temporal gap”, as implied by the term “imparfait de rupture” (‘imperfect of rupture’, , ; see also ). As we saw in Section 2, these adverbial expressions pertain to the group which relates times within the discourse. Due to the co‑occurring adverbial expression, the (first) imperfect form following it does not update the reference time itself. This is a possible basis for the interpretation found in the research literature that the imperfect still maintains its “co-preterite” function in this case (see, for instance, with further references). However, the verb does not depend on other verb forms to update the time. An example from our corpus study is (29). The typical narrative imperfect occurs in [4]. More specifically, the achievement verb llegaba (‘arrived’) is preceded by the sentence adverbial una hora más tarde (‘an hour later’). With respect to the temporal embedding, the tense-aspect forms in [3] and [6] have to be considered. They are both perfective (despegó, ‘took off’ and respondieron, ‘answered’). Sentence [4] may be considered part of the temporally updating structure of the main story line.
‘[1] Without stopping the engine, the pilot turned the plane around, [2] had it roll over the field for a while [3] and took off leaving behind a whirlwind of dust and straw. [4] An hour later, Onofre Bouvila arrived at the door of the house where he had been born; [5] a peasant now lived there with his wife and eight children. [6] To his questions they answered that the mayor lived in a new house, next to the church.’
In Figure 4, we visualize the temporal structure corresponding to clauses [3] to [6] of example (29). The eventuality in [3] is eventive in nature and has a corresponding reference time (R3 ). The typical narrative imperfect is preceded by a sentence adverbial which introduces a temporal gap with respect to the eventuality expressed in [3]. The anchoring relation between locAdv and R3 is not the standard sequential interdependency. The corresponding arrow is therefore dotted. Due to its lexically contributed specificity and the temporal gap, the location time of the adverbial (locAdv ) has an elevated prominence value. It is therefore represented in bold. As the narrative imperfect is part of the development of the narrative via its location time, which is temporally dependent on locAdv , a new reference time (R4 ) is introduced. Clause [5] introduces a state. The corresponding location time loc(s5) holds true at the previously introduced reference time (R4 ). The event contributed in sentence [6] again updates the narrative time. Via its location time (loc(e6)), which is anchored to the reference time introduced at the location time corresponding to the narrative imperfect, a new reference time (R6 ) is introduced into the structure.
At least in French, a typical narrative imperfect may be followed by one or more further updating imperfect tense-aspect forms (see ). These forms may then be said to contribute a temporal update “on their own”. Given the fact that other sub‑types (see below) may feature several imperfective forms in a row, it is probable that the Spanish typical narrative imperfect too can be followed by another updating imperfect. However, within our limited data set, this case did not occur. In reviewing the research literature, we only found examples which, although they do combine several verbs marked for imperfecto, upon a more detailed examination are not relevant cases. For instance, (30) shows a typical narrative imperfect in clause [1]. However, the nominal structure in [2] and, more importantly, clause [3] are elaborations of [1]. Thus, as we saw in § 3.1, they are rhetorically subordinate and do not advance the narrative time further.
‘[1] Five minutes later, the local team scored the equalizer, [2] another combination of Toñín and Eugenio, [3] and Toñín scored the three-three.’
The second sub-type we want to discuss here is termed the “imperfecto histórico” (‘historical imperfect’) in traditional descriptions of Spanish (see ). We also find the term “pretérito biográfico” (, ‘biographical preterit’, see § 3.1), highlighting the fact that it is sometimes used in the description of a person’s life. As we will see, its superficial structure is very similar to that of the typical narrative imperfect (see also ). It may be for this reason that it is not always distinguished from the typical narrative imperfect. For instance, presents several relevant examples, but classifies them as “imperfecto de ruptura” (, ‘imperfect of rupture’). However, the fact that, strictly speaking, there is no “rupture” is what distinguishes this use from the typical narrative imperfect. It also involves a temporal sentence adverbial. However, this adverbial expression of time pertains to the first group mentioned in Section 2 as it refers to a specific time point (or time span) in the possibly fictional extra-linguistic world (like en 1607, ‘in 1607’, as in example (31) below). Importantly, it does not place the embedded eventuality in a direct temporal relationship with respect to the preceding context, which would be the case in the rupture use. A typical context seems to be one where reference is made to an event which is historically important. This may also concern the status in a fictional world. Therefore, we suggest the term “imperfect introducing special events”. This is in line with , who calls it the “imparfait d’événement” (‘imperfect of occasion’). The distinctive property of being temporally detached is also mirrored by the fact that it may occur as part of the beginning of an episode (or formally, a paragraph).
Example (31) shows this case. With sentence [2] a new paragraph is started. Sentence [2], which begins with the sentence adverbial (en 1607, ‘in 1607’), contains a punctual use of the imperfective past tense aspect form (entraba, ‘entered’). As in the case of the typical narrative imperfect, the location time of this event depends on the location time corresponding to the adverbial expression (locAdv ). LocAdv is detached from previous reference times, but in line with the coherence of the text, it is located posterior to the previously given explicit adverbial location time corresponding to en el siglo xiii (‘in the 13th century’). However, paralleling the typical narrative imperfect, the punctual imperfect cataphorically takes part in the temporal structure, as the following event’s temporal location (descubrió, ‘discovered’) structurally depends on it. Thus, the location time corresponding to entraba (‘entered’), loc(e2), introduces a reference time R2 that functions as anchor for subsequent time points (at least loc(e3)). As the sentence adverbial en 1607 marks the beginning of a new episode, the corresponding location time (locAdv ) may potentially be maximally prominent in the domain. The relevant domain is the episode which the adverbial expression introduces.
Interestingly, our corpus data contain cases which involve several such structures in a row (see (32)). The research literature does not seem to mention this potential. While the basic functioning is not altered, the ascribed prominence potential is different. In this case, not all sentence adverbials introduce a new episode. Therefore, the corresponding location times (locAdv ) are not all maximally prominent; only the first maintains this potential. The example represents the beginning of a new chapter. It shows four imperfects introducing special events ([1], [3], [4], and [6]). Interestingly, in between these uses, the pretérito perfecto simple occurs ([2], [5]). Without much interpretive doubt, the clauses containing the uses of the imperfecto in question refer to even more critical historical information than the intervening sentences with verbs marked for the perfective past. Example (32) thus underlines the highlighting function of this use of the imperfecto. In fact, it may be said to show an “inverted” foreground-background relationship when accounted for in terms.
‘[1] On June 28, 1914, Franz Ferdinand was assassinated in Sarajevo together with his wife. [2] At the beginning of July, Vienna decided to present an ultimatum to Serbia. [3] On July 23 the note reached the Serbian government, with the indication that it had 48 hours to respond. [4] On the 25th Serbia replied that it accepted all the conditions except one. [5] Immediately, the Austro-Hungarian Empire broke off diplomatic relations [6] and on the 28th declared war on Serbia.’
It may be added that the imperfecto introducing special events may also be combined with other uses of the imperfective past. Again, example (33) occurs at the very beginning of a new chapter. Sentence [1] shows an imperfect introducing a special event. Sentence [2], by contrast, contains an instance of what we call the typical narrative imperfect. The main difference is that the adverbial expression in [1] refers to a text-external time point (el 2 de marzo de 1972, ‘on March 2, 1972’), while the one in [2] relates time points given within the text (trece meses después, ‘thirteen months later’). Therefore, their functioning as part of the temporal structure corresponding to the text is different.
‘[1] On March 2, 1972, the Pioneer 10 spacecraft launched from Cape Canaveral. [2] Thirteen months later, Pioneer 11 was launched. [3] Although its scientific mission was to investigate Jupiter and Saturn, both spacecrafts became famous for being the first to carry a message destined for extraterrestrials.’
In our French corpus study (see ), we found an interesting sub‑type, where several updating imparfait forms occur in literary texts without a sentence adverbial or other kinds of subordination. In this sub‑type, the imperfective tense-aspect form is used as a surprisingly good substitute for the perfective past, as it may occur throughout an episode (see ). Therefore, in this use, the reference times involved may acquire a wide range of prominence values (see ). However, this use appears to “be an individual preference of certain authors” ().
Our Spanish corpus data does not show the phenomenon as distinctively. While in the French data, it was quite versatile and not too seldom (see ), the few Spanish cases we have found are not unambiguous. In example (34), we may argue in favor of two updating imperfecto forms in [3] and [4]. The preceding verbs marked for imperfecto (and occurring in a main clause) are not updating, but describe a property (debía de medir, ‘must have measured’) and habitual behavior (se paraba, ‘stopped’; hacía frente, ‘faced’). However, in [3], a potentially single and punctual event is referred to by means of decidía (‘decided’). Via the corresponding location time (loc(e3)), a reference time is introduced into the temporal structure. In the case of the following main clause verb proseguía (‘continued’), co‑temporality seems to be ruled out. The indicator is inmutable (‘unshakably’), which expresses the reaction of the elephant with respect to the action of the rhinoceros, which consists of not reacting physically. This becomes even clearer given the fact that, in the directly preceding context (in [2]), the normal reaction of the elephant is described, which would be to pause (se paraba, ‘it would stop’). In [5] a non-adjacent but following event is expressed by means of a perfective past (vadeamos, ‘we waded’). The non-adjacency is rendered explicit adverbially (poco después del mediodía, ‘shortly after noon’). However, the location time corresponding to the wading event (loc(e5)) is still anchored to the reference time R4 introduced into the temporal structure via the location time corresponding to proseguía (loc(e4)). Given the fact that [1] and [2] are the beginning of the episode and present a description pertaining to the background, we might argue that the reference time R3 is the first reference time of the episode specifically introduced via an event, and therefore has an elevated prominence value. Along this line of argumentation, we would then be able to deduce that this use of the Spanish imperfect also allows for a diversification of prominence values. However, the prominence value of times in bordering positions with previous stative material has not yet been investigated in depth. Therefore, this insight is only preliminary at this point.
‘[1] The elephas indicus on which we were riding must have been about three and a half meters high [...]. [2] When it felt a threat approaching, [...] it stopped suddenly for a few seconds and faced it decisively [...]. [3] Some unicorn rhinoceros [...] running towards us at full speed, decided at the last second to pass by our side. [4] Unshakably, our elephant continued on its way while with its prehensile appendage it tore up bushes and reeds and peeled off fruit trees. [5] We waded across the Diflu River shortly after noon.’
In § 3.1, we mention the tendency of the imperfecto to occur more extensively and with possibly punctual readings in football match reports (see ). However, in the examples from the research literature, the use of the imperfect may be explained as the consequence of other grammatical or semantic factors or they may be too short to rule out such reasons (see also the short example in ). Thus, they cannot be determined unambiguously as featuring the updating uses we are interested in. However, in addition to the mentions they receive in the literature, the fact that there is a French counterpart to such usage begs the question of whether it occurs in Spanish. In French, the updating imperfect is pervasive in football match reports, as shown, for instance, in Labeau (, ) and investigated in depth in Egetenmeyer (, , accepted b). As notes, “the French imperfective may in fact appear as the only inflected verb form in entire newspaper articles on football”. Therefore, in French, the use of the updating imparfait found in football match reports is the most diverse sub-type (see ). It occurs with verbs of different Aktionsart, may but does not have to co‑occur with different kinds of adverbs, and the corresponding reference times may realize a varied set of prominence values (see Egetenmeyer , ).
Our corpus analysis of Spanish shows that updating uses of the imperfecto are in fact possible in football match reports. Although it cannot be said to be conclusive, as more data needs to be analyzed, it seems to indicate that, first, this phenomenon is generally less frequent than in French, and second, the diversity is most probably also not comparable to the case of French. Among the hits retrieved from the corpus, which seem relevant at first sight, many turn out to be either co‑temporal or part of rhetorically subordinate structures. We discuss this in § 3.1 and do not repeat it here. However, it might be a relevant factor for future studies to consider whether such embedded structures licensing the imperfecto might be more frequent in football match reports, as this would explain the impression that this tense-aspect form is especially frequent in this text type.
Example (35) is an excerpt from a football match report. It features an instance of the updating imperfecto in clause [3]. We include [1] to show that there is no co-temporality involved. In the passage, events in sequence are described. In [1], the team is left with one player less (se quedó, ‘was left’), in [2], a player gets a yellow card (se ganó, ‘earned’), and in [3] an unlucky shot is realized (estrellaba, ‘slammed’). The sequential relationship of [3] with respect to [2] is rendered explicit by the combination of the conjunction y (‘and’), and, more importantly, the adverb luego (‘then’). The updating function of the imperfect (estrellaba, in [3]) is thus not realized without indicators in the example, but the corresponding reference time R3 would be available for a subsequent anchoring relation. However, in the example, the following sentence [4] is introduced by an adverbial expression realizing a complex anaphorical relationship (see ). En ese decorado (‘in this scenario’) refers back to several previously presented events in which the team has been unlucky. We select this example because it also contains a typical narrative imperfect just two sentences after the updating imperfect mentioned. In [5], a sentence adverbial introduces a temporal gap (dos minutos después, ‘two minutes later’) and the main verb of the sentence is marked for the imperfective past (empataba, ‘equalized’). Thus, this kind of updating imperfecto is also possible in football match reports.
[4] En ese decorado […] Butragueño marcó el 0-1, al rematar de cabeza a placer un centro de Míchel desde la derecha. [5] Sólo dos minutos después, Juanito empataba, también de cabeza, al rematar una falta lanzada por Pizo Gómez. (CREA: La Vanguardia, 14/01/1994, Mendoza y Floro reciben un balón de oxígeno en última instancia)
‘[1] And, to make matters worse for Madrid, in the 32nd minute they were left with a man less due to the fully justified expulsion of Alkorta. [2] Shortly afterwards, in what seemed to be a general loss of temper on the part of Madrid, Míchel earned a card for protesting [3] and then Luis García slammed a ball against the base of the right post.
[4] In this scenario, Butragueño scored the 0-1 by an undisturbed header following Míchel’s center from the right. [5] Just two minutes later, Juanito equalized, also with a header, following a free kick by Pizo Gómez.’
We want to conclude the analysis with a discussion of the updating use of the imperfecto in the case of speech verbs. As noted, in her corpus analysis of radio and television data, shows that speech verbs followed by an adverb are a very frequent sub-type among the punctual uses of the imperfecto. Although there are competing factors, which may lead to the marking of a speech verb with the imperfective past tense-aspect form such as habituality or co‑temporality, these cases are relatively easy to find.
In one line of investigation, we considered literary data. There, the updating property becomes especially evident when the form is used to mark a speech verb in the representation of a dialogue. This is the case in example (36). In this extract about the finding of a stray lion cub, [3] represents the direct speech event and the anchoring speech verb involved. The speech verb (comentaba, ‘commented’) may be understood to mark a single speech event following the arrival of the ego narrator (and a second person) at the transfer point. The verb is part of the narrative progression. Therefore, there is a corresponding reference time in the temporal structure (R3 ). More specifically, in line with and our interpretation in , with reference to ), the speech verb in [3] is co-specified by the direct speech event. The subsequently introduced longer string of direct speech ([4]) follows the speech event of [3]. It has a certain location time (loc(e4)) via which a new reference time (R4 ) is introduced (see for a parallel analysis focusing on free indirect discourse). Importantly, loc(e4) is temporally anchored to R3 . There are several arguments favoring an independent speech event of [4] with respect to [3]. Syntactically, [4] is not embedded under the speech verb in [3]. The line break is a formal indicator. Finally, we might take lexical and semantic properties as an additional indicator as [4] is not a comment (see comentaba, ‘comented’ in [3]) but narrates a set of events.
[2] Por el camino, acabé de contarle todo lo que sabía.
[3] ―¡Qué sorpresa nos hemos llevado! ―comentaba el matrimonio catalán que acababa de salvar a Misuko.
[4] ―De repente, hemos visto un cachorro de león cruzando la carretera. Al principio, cuando paramos no nos atrevimos a acercarnos. Estaba asustado. Gruñía. Después, vimos un cartel con las palabras “parking” y “zoo”, y comprendimos.
[5] En ese instante, también comprendí yo. (CREA: Pardo de Santayana, El beso del chimpancé. Divertidas e insólitas historias de la vida cotidiana en un zoo, 2001, p. 203)
‘[1] “Misuko is at the ticket office,” I told Maribel. [2] On the way, I finished telling her everything I knew. [3] “What a surprise!” commented the Catalan couple who had just saved Misuko. [4] “Suddenly, we saw a lion cub crossing the road. At first, when we stopped, we didn’t dare to approach it. It was frightened. It was growling. Then, we saw a sign with the words “parking” and “zoo”, and we understood.” [5] At that moment, I also understood.’
We also included queries focusing on newspaper data. The results corroborate that the pattern of a speech verb marked for imperfective followed by an adverbial expression is frequent (see ). However, an unambiguous sequence appears to be much less frequent than one might expect. A possible example is (37). Sentence [2] contains the speech verb in question, which is marked for the imperfective past (opinaba, ‘said’). The information conveyed by the indirect speech is well integrated into the text and we may assume that the referent cited in the main clause stands in a relevant relationship with respect to the context. This is an important difference with respect to example (38) below. The main verb (sustituyó, ‘replaced’) in sentence [1] is marked for perfective past. Although the events referred to in [1] and [2] are not temporally adjacent, they follow one another. Therefore, we may classify the case as one of a temporal update.
[2] Al Ejército no le gusta que le metan en asuntos que no son de su competencia […], opinaba ayer el diputado Gueorgui Manchulenko […]. [3] Según Manchulenko, en la cúpula de los servicios de seguridad está el primer vicepresidente, Satiuk, protegido directamente por Víktor Medvedchuk, el jefe de la Administración presidencial. (CREA: Internacional)
‘[1] Kuchma thus replaced Yegueni Marchuk, a veteran of the security services, who was not trusted. [2] The army does not like to be involved in matters that do not fall within its competence, said deputy Georgy Manchulenko yesterday. [3] According to Manchulenko, at the top of the security services is the first deputy chairman, Satiuk, who is directly protected by Viktor Medvedchuk, the head of the presidential administration.’
However, many of the hits from our queries actually designate what may be called “single speech events”. This is exemplified in (38). Sentence [1] presents an assessment referring to the moment of speech. The main verb is marked for present tense (es, ‘is’). Sentence [2] adds an appeal which relates to the situation presented in [1] (and before), but is not dependent on it. Crucially, there is no sequential temporal relationship between [1] and [2]. Rather, [2] conveys further relevant information. This is similar in the case of sentence [3], which also includes a verb marked for imperfecto. Although some of the instances of this kind can be classified as elaborations in the sense presented in § 3.1, the relationship is not always the case, as shown in (38). When analyzing the temporal structure of such examples, we find that the speech verbs marked for the imperfective past cannot be said to update time. Their truth value is asserted for a specific time point, often rendered explicit by the co‑occurring adverbial expression, and we could say that via their location time a reference time is introduced. However, in examples of this kind, the respective time points do not enter into a clear temporal relationship with other time points contributed via eventuality expressions in the same domain. Simply put, they do not form part of the narrative sequence in a broad sense of the term. Often, there are no reciprocal temporal anchoring relations to be found in the discourse segment, as the only kind of anchoring is the relationship with respect to the speech time (n). By contrast, the temporal structure corresponding to the inner-textual relationships may be incomplete. An alternative and promising way to employ a discourse structural analysis is to turn to the textual development of such passages and focus on argumentative relationships and a “furnishing update”, which accounts for certain ways of adding “stative” information (see Egetenmeyer accepted a).
‘[1] And Spain, with 48 cases, is the third European country most affected by AIDS in children, all of them linked to the mother’s drug dependence. [2] Although Spain has one of the lowest rates of heterosexual infection in the EC ― 59 cases ―, García Vargas appealed yesterday for people to take precautions in sexual contacts with people who are not certain not to be carriers. [3] Health sources considered the World Health Organization’s message disseminated yesterday to be excessive.’
In the present contribution, we have dealt with a set of uses of the Spanish imperfective past tense-aspect form, among which the most renowned is called the narrative imperfect. By focusing on a specific property, namely the potential to update time, which has not yet been examined thoroughly in this realm, we have elaborated and delimited new categories of uses of the imperfect. They are analyzed as different cases of the updating imperfect (imperfecto actualizador). We have based the delimitation on further, temporally relevant properties of the uses, such as their embedding under a sentence adverbial or the co‑occurrence of an adverb and the potential diversification of prominence values of the corresponding reference times. The reference times, which are an essential component of temporal discourse structure (see ), allow us to properly operationalize the notion of temporal update.
We have shown that not all examples discussed in the research literature as cases of “imperfecto narrativo” do update the reference time. We find not only cases embedded under a habitual operator, but also co‑temporal ones. A factor perhaps not yet appreciated enough is rhetorical subordination, which may also inhibit temporal development. The use sometimes called “imperfecto de ruptura”, which basically is an imperfect preceded by a sentence adverbial expressing a temporal interruption, is the one most frequently termed narrative imperfect. We show that this is but one use among several, although a very salient one. We call it the “typical narrative imperfect”. We discuss a second sub‑type, which also follows a sentence adverbial, but which designates a specific time point and not a temporal gap. In these two cases, the location time of the adverbial may have an elevated prominence value. On the other hand, the reference time introduced via the location time corresponding to the eventuality expressed by means of the verb marked for the imperfective past is normally equal by default. Next, we discussed several cases which are not embedded under a sentence adverbial. More specifically, the third case occurs in literary data. There, the verb marked for the imperfective may update time without an adverbial and the prominence values of the corresponding reference times may show a certain variance. Fourth, the updating imperfect also occurs in sports reports (newspaper data), although this takes place less frequently than expected. Fifth, we accounted for speech verbs marked for the imperfecto and their potential to update the reference time.
In comparison to French (see our corresponding previous study in ), our corpus analysis suggests that updating uses of the imperfective past are less pervasive in Spanish. This is also underlined by . Across the different sub-groups, the French imparfait has a stronger tendency to mark several verbs in a row, all of which may contribute a temporal update, and also tends to have more verbs in a row (see ). However, the imperfecto may also have further updating uses not covered in the present contribution. It is a vast field worth exploring further.
The research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) as part of the Collaborative Research Center 1252 Prominence in Language, Project-ID 281511265, which is gratefully acknowledged.
We thank Mary Chambers for proofreading, and Viktoria Henn and Anja Doveren for practical assistance. We thank Susana Azpiazu Torres for initiating this anthology. She also gave us feedback on an earlier version and discussed several examples with us, for which we are very thankful. Finally, we thank two anonymous reviewers for their insightful comments and suggestions which helped improve the paper further. All remaining errors are ours.
2
3
Azpiazu Torres, S. (2015): “El pretérito perfecto compuesto y el imperfecto narrativo en la prensa audiovisual peninsular”. Moenia 21, 23‑39. DOI: https://doi.org/10.15304/m.v21i0.2833.
4
5
6
7
8
9
Bello, A. (1883 [2012]): Obras completas de Don Andrés Bello. Volumen 4. Gramática de la Lengua Castellana. Edición hecha bajo la dirección del Consejo de Instrucción Pública en cumplimiento de la lei de 5 de setiembre de 1872. Edición digital a partir de Santiago de Chile. Impreso por Pedro G. Ramírez. Alicante: Biblioteca Virtual Miguel de Cervantes. http://www.cervantesvirtual.com/nd/ark:/59851/bmc4t6v6.
10
Berthonneau, A.-M. & G. Kleiber (1999): “Pour une réanalyse de l’imparfait de rupture dans le cadre de l’hypothèse anaphorique méronomique”. Cahiers de praxématique 32, 119‑66. http://praxematique.revues.org/2853.
11
12
13
14
15
17
Bres, J. (1999): “L’imparfait dit narratif tel qu’en lui-même (le cotexte ne le change pas)”. Cahiers de praxématique 32, 87-117. https://journals.openedition.org/praxematique/2852.
19
Brucart, J. M. (2003): “El valor del imperfecto de indicativo en español.” Estudios Hispánicos 6, 193-233. http://filcat.uab.cat/clt/publicacions/reports/pdf/GGT-03-1.pdf.
20
21
23
24
25
Egetenmeyer, J. (2020): “Temporal prominence demotion in updating imparfait uses in L’Étranger”. En E. Corre, D. Thành Do-Hurinville & H. Linh Dao (eds.): The Expression of Tense, Aspect, Modality and Evidentiality in Albert Camus’s L’Étranger and Its Translations: An empirical study. Amsterdam / Philadelphia: John Benjamins Publishing Company, 118‑42. DOI: 10.1075/lis.35.07ege.
26
27
28
29
30
31
32
33
35
36
37
38
40
Gutiérrez Araus, M. L. (1996): “Sobre los valores secundarios del imperfecto”. En F. J. Grande Alija, J. Le Men, M. Rueda Rueda & E. Prado Ibán (eds.): Actuales tendencias en la enseñanza del español como lengua extranjera II: actas del VI Congreso Internacional de ASELE (León, 5-7 de octubre de 1995). León: Universidad de León, Servicio de Publicaciones, 177‑85.
41
42
43
45
46
47
48
49
50
51
52
53
Labelle, M. (2003): “Events, states and the French imparfait”. En Y. Roberge & A. T. Pérez Leroux (eds.): Romance Linguistics: Theory and Acquisition: Selected Papers from the 32nd Linguistic Symposium on Romance Languages (LSRL). Toronto, April 2002. Amsterdam / Philadelphia: John Benjamins, 165‑80.
54
55
56
57
58
60
61
62
63
64
68
69
Rojo, G. (1974): “La temporalidad verbal en español”. Verba. Anuario Galego de Filoloxía 1, 68-149. http://hdl.handle.net/10347/2724.
70
71
74
75
77
78
80
84
El País. https://elpais.com/diario/1997/05/10/espana/863215204_850215.html.: Accessed: 29.04.2023.
85
CREA = Real Academia Española: Banco de datos (CREA) [en línea]. Corpus de referencia del español actual. http://www.rae.es. Accessed: 29.04.2023.
[1] This is a somewhat superficial but helpful approximation with respect to the set of phenomena we deal with in this paper. However, in the literature there is some discussion on the matter (see fn. 3).
[2] For a recent extensive overview of publications dealing with the narrative imperfect the reader is referred to .
[3] Many scholars concerned with the Spanish narrative imperfect mention a parallelism of the narrative imperfect with respect to the perfective past or address the occurrence of the former in a context where the latter would normally be expected. Among them are, for instance, , , , , , , , , . Some other scholars assume that the narrative imperfect maintains its original sense which leads to a partly different reading of the examples in question. We could mention here, for instance, Azpiazu Torres ( with further references, ), , , (sometimes cited as ), and . presents an extensive literature overview involving a relevant distinction, namely, whether the accounts operate on an aspectual or a temporal interpretation of the narrative imperfect.
[4] Relative clauses are nominal in nature (see , with reference to ). Therefore, they have a certain tendency to convey descriptive material. In terms of , this pertains to the background of a narrative.
[5] By contrast, the state expressed in [3] cannot hold true only at the time of sentence [1] (see ).
[6] There are, however, also other understandings of the “adverbios de marco” (‘framing adverbs’). For instance, García Fernández ( with reference to ) defines them in pure content terms. In his classification, “adverbios de marco” refer to an extended time span and are opposed to the non-extended “adverbios de punto” (‘punctual adverbs’) (see ). refines another point. According to her, the extension is not contributed by the adverbial expression itself but by the clause predicate (see ). We do not agree fully with her on this point, because although the expressions she discusses may be under-specific with respect to an eventuality’s extension, they actually do have lexical meanings expressing different extensions (e.g., a las dos, ‘at two o’clock’, vs. esta tarde, ‘this afternoon’, see ). However, this does not affect the relevance of the three groups she distinguishes, namely, “complementos temporales durativos, los de fase y los de frecuencia” (, ‘durative temporal complements, those of phase and those of frequency’). She thereby refines second group. His first group is partly based on a discourse-functional and partly on a referential scheme and consists of deictic, anaphoric, and calendrical adverbial expressions (see ).
[7] A similar distinction is discussed in with reference to ) in the context of the pluperfect’s two readings as expressing an event or a post-state.
[8] In fact, neither nor actually focuses on the second verb which is of interest to us. Rather, they focus on indirect speech and take the example to show the functioning of tense in embedded clauses.
[9] The example presented in NGLE, however, is more abstract than ours. In (i), the macro event is only implied, namely as the whole day, while [1] and [2] occur one after the other (see ).
[10] Given the lack of previous context, the status cannot be determined unequivocally. If decía (‘said’) pertained to a sequence of events, an own reference time would be associated with it. However, if it elaborated a previously given speech event (in the sense that the sentence in the example expresses what was actually said), the corresponding location time would include the previously given reference time. This second case may be said to be less probable given the very detailed subject NP (el presidente del Eurogrupo, J. C. Junker, ‘the president of the Eurogroup, J.C. Junker’). Also ayer (‘yesterday’) may be an indicator countering the second reading.
[11] In this specific case, we have not been able to retrieve further preceding context. We thank Susana Azpiazu Torres (p.c.) for pointing out to us that the example is the title of a news item. Interestingly, as she tells us, the news item continues with a focus on unemployment in Spain instead of talking about Junker (Susana Azpiazu Torres, p.c.).
[13] An alternative set of situations would be that the player first scores a goal, and then, after running over to the stand, actively interacts with the fans to motivate them to shout and applaud. Then there would be two distinct events.
[14] It is not fully clear if himself refers to the level of assertion. It seems to be left open in the paper. However, we take this to be the crucial level for the analysis of the temporal discourse structure.
[15] By contrast, a backgrounded reading is improbable, and a habitual reading seems to be excluded.
[16] Interestingly, languages with a specific form expressing progressivity may allow its use in such contexts. This would be the case of English As I was saying, used to re-introduce a certain topic discussed earlier in a conversation. As points out, the English progressive can also have an evaluative use in dialogue. It should be noted that we do not deduce the status of the Spanish example from the translations to English.
[17] With respect to the verb in [3] marked for imparfait, a progressive reading may also be possible, but is less probable.
[18] The focus Bertinetto () places on the wider extension of the referred time of the imperfective verb, however, might be viewed critically. In terms of assertion (see fn. 6 above), it is not unequivocal if the eventuality in [3] really is extended or if this is simply inferred.
[20] It shares certain parallels with the “presente histórico” (‘historical present’), a tense use mentioned frequently in the research literature (see , , ; etc.).
[21] The chapter is 4 La Primera Guerra Mundial (‘4. The First World War’), which starts right away with sub-section 4.1 Las Condiciones Políticas (‘4.1 The political conditions’) (CREA: Quintanilla & Sánchez Ron, Ciencia, tecnología y sociedad, 1997, p. 73).
[22] The chapter is called Mensaje en una botella (CREA: Sabadell, El hombre que calumnió a los monos, 2003, p. 135, ‘message in a bottle’).
[23] In § 3.2, we discuss the difficulty that longer passages involving the imperfecto as a predominant past tense-aspect form may actually have a habitual reading. It may be necessary to consider a very extended context to determine this status.
[24] The indefinite article algún (‘some’) would also be compatible with a habitual reading. However, it does not induce this other reading unambiguously either.
[25] Otherwise, this could also be a trigger of the imperfective past tense-aspect form of decidía (‘decided’) in [3], because the trigger may always have a reciprocal effect between two (or more) related verb forms.
[26] She uses the term “imperfecto narrativo” (see Azpiazu Torres 2015: 29). As we use the term more restrictively, we do not follow her in the terminology.
