Chapter Tense and Aspect

by Östen Dahl and Viveka Velupillai

Please note: This is an introduction to Chapters 65 (Perfective/Imperfective Aspect), 66 (The Past Tense), 67 (The Future Tense), and 68 (The Perfect).

General introduction

The following four maps concern the role of tense and aspect in the grammars of the world’s languages. Traditionally, tense and aspect are seen as grammatical categories of verbs. Indeed, insofar as verbs display morphological variation in languages, there is a strong tendency for such variation to include inflectional differences between forms that reflect temporal and aspectual distinctions, such as the difference between the present tense form sings and the past tense form sang in English. However, some modifications to the identification of tense and aspect with grammatical categories of verbs must be made. To begin with, it has long been recognized that periphrastic constructions, such as the English Perfect (I have sung) and Progressive (I am singing), are employed in functions similar to those of inflections. Furthermore, tense and aspect do not always present themselves as separate and neatly delineated categories, as we shall now argue.

Among the multitudinous definitions of tense and aspect in the literature, we may cite those given in Comrie (1985: 6): tense is grammaticalisation of location in time, and aspect is “grammaticalisation of expression of internal temporal constituency” (of events, processes etc.). Thus defined, the two categories are conceptually close in that both deal with time. But they may also be interwoven in grammatical systems in that one and the same grammatical form may combine temporal and aspectual elements in its semantics. For instance, the very common and central distinction between imperfective and perfective verb forms would appear to be a straightforward example of aspect; yet it typically involves not only aspectual notions but also restrictions on temporal reference, in that perfective verb forms are usually taken to refer to past events. This fact has a cognitive basis: we may say that the prototypical uses of perfectives coincide with the default view of an event as a completed whole. But normally such a perspective is possible only if the event is wholly in the past. Therefore, in languages which do not mark the imperfective-perfective distinction, verbs that are typically event-denoting such as ‘arrive’, ‘close’, and ‘kill’ may be interpreted as referring to the past in the absence of any marking to the contrary. Verbs that typically denote states such as ‘sleep’, on the other hand, tend to be interpreted as referring to the present. (In addition, verbs are sometimes indeterminate or ambiguous between referring to a present state or to the event in the past that gave rise to it, e.g. German er ist gestorben ‘he is dead’ or ‘he died’).

What has been said here about tense and aspect may also be extended to the closely related category of mood. An alternative to seeing tense, aspect and mood as grammatical categories in the traditional sense is to regard tense-aspect-mood systems as wholes where the building-blocks are the individual tenses, aspects, and moods, such as the Past and the Progressive in English. These will be referred to as grams, and it is assumed that on the cross-linguistic level they represent a restricted set of gram types. This is the approach that will be taken here. In the following, the names of language-specific grams will be capitalized, to distinguish them from gram types, which are written in lower-case characters.

The term grammaticalization, as used in the definitions from Comrie cited above, refers to a synchronic property characterizing a notion (semantic category) if and only if it is reflected in or determines the use of grammatical items. Obviously, both location in time and the internal constituency of events may receive linguistic expression in multifarious ways, which are not, however, relevant to the notions of tense and aspect as defined here, as long as they do not play a systematic role in the grammar. For instance, temporal adverbials such as yesterday or in 2001 are also used to locate events in time, but they differ from grammatical tense in fundamental ways. While temporal adverbials are used only when they express information which is relevant to the particular intended message, the use of tenses is guided by general principles that often make the choice of a certain tense obligatory and that make the use of a tense morpheme obligatory even if the information it carries is redundant. Thus, in a sentence such as Last year I bought a new car the choice of a tense other than the simple past would make the sentence anomalous, although the information that the event took place in the past is expressed unambiguously by last year.

In the case of aspect, the delimitation of what is grammatical(ized) is more difficult, and further confounded by the lack of a consistent terminology. In some research traditions, the term “aspect” is used to refer to a wide domain of phenomena, including many that are not manifested as grammatical distinctions (cf. for instance Verkuyl (1971), Tenny (1994)). Others make a strict distinction between aspect as a grammatical phenomenon and “Aktionsart” as pertaining to lexical or purely notional (semantic) categories (this tradition goes back to Agrell (1908)). However, among those who make a distinction between aspect and Aktionsart, there is no unanimity as to how the latter term should be used. Especially in Slavic linguistics, the term “Aktionsart”, or its counterpart in other languages (such as sposob dejstvija in Russian), is used for phenomena that straddle the borderline between grammar and lexicon, notably various derivational processes by which verbs with specific aspectual meanings may be created. For instance, from the simplex Russian verb spat’ ‘to sleep’ we may obtain the verb pospat’ ‘to sleep for a while’ by adding the “delimitative” prefix po-. With a more liberal definition of aspect, such processes would be called “derivational aspect”. Although there is considerable variation as to the extent to which derivational aspect is elaborated and used in languages, relatively little work has been done with the aim of systematizing this variation. In these chapters, we have therefore chosen to focus on grammatical aspect in the narrowest sense, where typological research has already advanced far enough to make it possible to define mappable parameters. However, some areal tendencies may be mentioned here. Derivational aspect (Aktionsart) appears to be particularly well developed in many indigenous North American languages; eastern Europe could be another example of an area where these phenomena are well represented.

Today, “grammaticalization” is probably most often understood in a diachronic sense—as the development of grammatical marking over time, typically from lexical sources. The grammaticalization paths for tense and aspect are relatively well known, at least in the languages for which there is sufficient information about earlier stages (which is certainly not a representative sample of the world’s languages). It has thus, for instance, been established that both pasts and perfectives may arise from perfects, whereas imperfectives often develop out of progressives (see Bybee et al. 1994: 51-175); iteratives and similar constructions as well as futures may come from a number of different sources, such as verbs of volition, obligation and motion (Bybee et al. 1994: 243-280). There is a strong correlation between the way in which a tense-aspect gram is expressed—whether it is inflectional or periphrastic—and how far it has advanced on its path of grammaticalization. Thus, perfects and progressives are overwhelmingly periphrastic, whereas past and perfectives are more prone to be inflectional. 

It goes without saying that it is impossible to encompass the richness of the tense and aspect systems of the world’s languages within four maps. We have been forced to neglect a number of gram types, most of which have interesting geographical distributions; among these are habituals, iteratives, frequentatives, dedicated narrative forms, resultatives and experientials. We have also neglected tense and aspect marking in embedded contexts, as well as the interplay between tense-aspect and other categories, such as negation (many languages have special tense-aspect forms in negative contexts, cf. Chapter 114).

Tense and aspect are notoriously difficult categories to describe adequately, and the treatment in grammars is often problematic, especially if one wants to use it for cross-linguistic comparison. As far as possible, we have tried to apply consistent criteria in classifying tense-aspect phenomena. For this reason, our interpretations sometimes differ from those found in grammars. The reader should thus not be surprised if a language is classified in an unexpected way. 

In compiling the data for the maps on tense and aspect, we have drawn upon two earlier large-scale typological surveys, namely those presented in Dahl (1985) and Bybee et al. (1994). The sample is the same for all four maps and contains all languages in the basic WALS 100-language sample except one (Mezquital Otomí), for which it was not possible to get adequate information. Discussions of areal tendencies in tense-aspect systems based on the sample in Bybee et al. (1994) are found in Dahl (1995) and Dahl (2000); the present sample is about three times as large, however, and although earlier findings are basically confirmed, we are now able to see the general pattern much more clearly.