Navigation – Plan du site

The Parameters of Aspect for Oḍia

Bibhuti Bhusan Mahapatra
p. 21-46


On suppose généralement que les aspects situationnels (les aktionsarts) sont marqués lexicalement sur le prédicat. En Oḍia, pourtant, ils peuvent être exprimés grammaticalement au moyen d’un ensemble fermé d’auxiliaires. Concernant les aspects marqués grammaticalement (les aspects de point de vue), l’Oḍia possède un marqueur progressif mais n’a pas de marqueur perfectif ou imperfectif visible. Le marqueur habituellement appelé « perfectif », qui sert aussi de marqueur de « participe conjonctif », reçoit ici une analyse unifiée comme marqueur de linéarisation verbale. Notre analyse montre également qu’en Oḍia, la perfectivité prend la forme d’une séquence qui consiste en un verbe événementiel plus un auxiliaire copule, c’est-à-dire un événement suivi d’un état. Nous faisons remarquer qu’à travers les langues, le morphème progressif ne peut pas recevoir d’interprétation en l’absence de traits de temps et de monde. Au contraire, le morphème imperfectif efface ces traits. En somme, ce travail propose que « l’imperfectif » et « le présent simple » sont des structures alternatives paramétrées qui ne valident pas les traits de temps et de monde. Bien que la forme progressive porte invariablement un temps grammatical, sa valeur aspectuelle varie selon les langues. En Oḍia, sa valeur est [+Dynamique] mais en anglais sa valeur est [−Télique]. Les valeurs [+Dynamique] ou [−Télique] constituent un choix paramétrique des langues.

Haut de page

Texte intégral

Oḍia (in earlier spelling ‘Oriya’) is the official language of the State Oḍisha (in earlier spelling ‘Orissa’) in eastern India.

I thank la Maison des Sciences de l’Homme (MSH, Paris) for ha-ving granted me a short-term fellowship in 2004 to begin the initial phase of research on the facts presented in this paper. For helpful suggestions at va-rious stages of the development of this article I thank Jacqueline Guéron, Bridget Copley, K. A. Jayaseelan and R. Amritavalli.

1. Preliminaries


2The following abbreviations are used in this text:


Ablative case


Imperfective aspect


Accusative case


Locative case








Past tense




Perfective aspect




Present tense


Completive auxiliary






Progressive aspect


Definite Marker


Stative verb or auxiliary




Verb linearizer



3Studies on aspect recognise that the aspectual information of a clause is indicated in two distinct domains. The predicate is lexically marked for aspectual properties (Aktionsarts); and the clause grammatically indicates aspectual features by aspect morphemes.

  • 1 Smith (1997: 19) takes ‘static’ and ‘dynamic’ as equipollent terms; thus, their privative oppositio (...)

4Vendler (1967) brings into focus the lexically marked aspect features of verbs and accordingly classifies them as ‘states’, ‘activities’, ‘accomplishments’ and ‘achievements’. Smith (1997) adopts Vendler’s verb classes as ‘situation types’ and proposes semelfactive as another ‘situation type’. The aspectual features that motivate Vendler (1967) to identify different verb classes are given in Smith (1997: 19-20) with a few binary terms as follows:1

5Smith (1997) contrasts lexically marked ‘situation types’ with grammatically marked aspect features which she calls ‘viewpoint’ aspects.

6Carlson (1977) classifies predicates as stage-level and individual-level. He considers them as semantic primes, without exact grammatical correlates. However, he (1977: 168, 170) just suggests that perfective and progressive are incompatible with individual-level predicates.

7In this paper, I assume that stage-level and individual-level features along with the Vendlerian aspect features lexically form the aspectual features of the predicate; they are projected onto the aspect structure by the predicate and are checked by appropriate grammatical features.

1.1. Grammaticalized situation-types

8Although situation types are assumed to be inherent to the lexical meaning of the predicate, interestingly, in Oḍia some of them can be expressed with grammatical auxiliaries; such auxiliaries can combine with the main verb to make its aspectual sense grammatically explicit. Consider the auxiliary rah (stay) in the following sentences:

  • 2 The state that the auxiliary rah (stay) denotes is a positional state which can be contrasted with (...)

9In (2) the auxiliary rah goes well with the stative verb so (sleep); but in (3) it does not go with the activity verb coba: (bite). Thus, (2)-(3) illustrate the stative function of rah.2 Now, consider the behaviour of the auxiliary la:g (stick/engage/continue) in the following sentences:

10In (4) the aspectual auxiliary la:g does not go with the stative verb so (sleep); but, in (5) it goes well with the activity verb coba: (bite). Thus, (4)-(5) illustrate the activity function of la:g. Further, consider the behaviour of the auxiliary ca:l (walk/move/continue) in contrast with la:g:

11In (6) the verb pel (push) indicates an activity; but in (7) the unaccusative, poḍ (burn) does not indicate an activity, but simply a process. For a verb to have the activity sense it has to select an agent. Unaccusatives, which do not select an agent, do not have the activity sense. The auxiliary la:g in (6) goes well with the verb pel (push), which denotes an activity; but in (7) it does not go with the unaccusative poḍ (burn), which does not denote an activity, but a process. However, as (8) illustrates, the auxiliary ca:l (continue) goes well with the unaccusative poḍ (burn), which indicates a process. Thus, (7)-(8) illustrate the process function of ca:l.

12Verbs which select la:g (activity) can alternatively select ca:l (process). The fact is illustrated in the following sentences:

  • 3 These Oḍia facts simply confirm Davidson’s (1967) view that an action (activity) can be seen as an (...)

13The possibility of ca:l (process) being selected in the place of la:g (activity), as illustrated in (9)-(10), implies that activities can be viewed as processes. However, the alternative view, that is, the activity view of a process is not available in the language,3 as evident from the behaviour of the unaccusative illustrated with (7).Although the language allows the process reading of the activity it still keeps them grammatically distinct.

14There is another auxiliary in Oḍia, namely sa:r (complete/finish). It does not go with states or processes or activities; but it goes with accomplishments, as illustrated below:

15On the surface, the auxiliary sa:r (finish) has a completive function in (14); but, the completive morpheme grammatically fuses the dynamic (process) feature of the event with its end point (telos). So, it can only select situations which have a dynamic (process) part followed by its telos; and, none but the accomplishment type has such features. Therefore, the auxiliary sa:r (finish) is selected only with accomplishments.

16On the basis of the facts given so far, the following are the ‘situation-type’ auxiliaries and their related aspect features in Oḍia:

17Although ‘achievements’ and ‘semelfactives’ are assumed as distinct situation types, Oḍia does not show any overt aspect morphemes for them. This simply shows that a category which is semantically postulated is not necessarily grammaticalized in a given natural language.

  • 4 The VL (Verb Linearizer) -i in (16) is a default connector of verbs in Oḍia. I will come back to it (...)

18Although, in many languages, the situation type features are just inherent to the lexical meaning of the predicate, they can be combinatorial in Oḍia. Their combinatorial nature in Oḍia has to be taken as a syntactic fact. For example, the sentence (6) can be given the following structural representation (omitting here the higher functional categories, namely, viewpoint aspect, copular auxiliary, tense and agreement to which I will return):4

  • 5 Following Chomsky (1995) I assume that a lexically marked feature is checked (interpreted) in a con (...)

19In (16) and in the rest of this paper, I adopt the right-branching structure for Oḍia, and assume that the main verb moves higher up to incorporate the verbal features; and by doing so, it also checks its inherent aspectual features.5

2. Viewpoint aspects

20As far as I know, there is no study of the Oḍia aspect system which theoretically assumes the relevance of situation types. However, studies dealing with the verb morphology in Oḍia usually identify two aspect morphemes (cf. Nayak (1987: 34), Mohanty (1992: 46-47) and Sahoo (2001: 78-79)). Consider the following sentences:

21The verb forms in (17) and (18) provide the minimal contrast between the morphemes -u and -i; and so do the verb forms in (19) and (20). On the basis of the contrasts between the verb forms such as in (17) and (18), and (19) and (20), -u is analysed as the progressive and -i as the perfective aspect markers. Smith (1997) takes grammatically marked aspects such as progressive, imperfective and perfective as viewpoint aspects. This straightaway identifies the progressive -u and the perfective -i as the viewpoint aspect markers.

2.1. Ambiguity of the progressive

22The progressive has ambiguous functions across languages. It can behave as a frequentative in both Oḍia and English if the predicate bears atelic process reading. Consider the following Oḍia sentence and its English gloss:

23In (21), the temporal adverb which coerces an atelic process reading of the basic accomplishment sense of the verb kha: (eat) gives frequentative meaning to the progressive -u.

2.1.1. The Progressive goes with processes

24In fact, the progressive -u goes with clear atelic processes. Consider the following sentences:

2.1.2. The Progressive goes with activities

25Activities conceptually do not have a telic feature either. The morpheme -u also goes with them. Consider the following sentence:

26To sum up, the progressive -u in Oḍia goes with accomplishments (e.g. (17) and (19)), with processes, (e.g. (21)-(22)) and with activities, (e.g. (23)). Thus, it seems to bear an aspectual function common to all of them.

2.1.3. The Progressive does not go with states

27However, the progressive -u does not go with verbs which have stative sense. Consider the following sentences:

28The verb dhar (hold/catch) is ambiguous in Oḍia. The sentence (24) illustrates that its stative sense is not available with -u. The sentence (25), in contrast, illustrates that its stative sense is available with the aspect morpheme -i.

29If -u occurs with a stative verb, the verb acquires either a process or an accomplishment reading. For example, the verb dhar (hold/catch), which has a stative (hold) reading with -i in (25), has a process or an accomplishment (catch) reading with -u in (26):

30However, (24)-(26) very well support the analysis that the progressive -u is inconsistent with the stative sense of a verb. They also illustrate that the two different senses of dhar, namely hold and catch, which are aspectually distinguished in Oḍia are lexically realized as different verbs in English.

31Some other Oḍia verbs whose stative sense is separable from their Process/Accomplishment sense with the choice of viewpoint aspects are: cã:h (be awake/look (at)), ja:ṇ (know/realize/understand/learn), paḍ (lie/fall)...

2.2. Progressive as [+Dynamic]

32As we saw, the progressive -u goes with accomplishments, processes and activities, but not with states. In fact, it grammatically bears the [+Dynamic] feature, which is common to accomplishments, processes and activities; and it is in binary opposition with the [−Dynamic] feature, which marks the state. I will return to some broader implications of the grammatical aspects triggering [±Dynamic] feature contrast later.

2.3. Problems with the perfective reading of -i

33Typically, the position immediately preceding the copular auxiliaries is identified as the position where the contrast between the perfective -i and the progressive -u holds. That is the only position where -u can occur; and -i in contrast with -u in that position is interpreted as the perfective aspect, as illustrated in (17)-(20). However, this analysis of the morpheme -i faces a number of problems in different contexts:

2.3.1. When governed by the copular auxiliary

34Consider the following sentence:

35In (27), the verb so (sleep) refers to the lion’s sleeping state; but, there is no perfection of the event implied there. Thus, the usual perfective analysis of -i in the position immediately to the left of copular auxiliaries is not available with states.

2.3.2. When governed by the modal auxiliary

36Further, consider the following sentences:

37(28)-(29), which disallow the progressive -u in the position immediately to the left of the modal auxiliary, illustrate that such positions are not valid positions for viewpoint aspects. Nevertheless, the morpheme -i does occur in such positions, as illustrated with (30)-(31) below:

  • 6 Sahoo (2001: 85-87) labels the -i occurring just before modal auxiliaries and light verbs as simply (...)

38Therefore, it cannot be assumed that -i has any aspectual sense in contrast with the progressive -u in the position before modal auxiliary in (30)-(31). Thus, the typical analysis of the morpheme -i as the perfective marker is contradicted in such positions.6

2.3.3. When governed by situation type auxiliaries

39Further, if we consider the occurrences of the morpheme -i with situation type auxiliaries in (2), (5), (8) and (14), the morpheme -i has no perfective sense in any position in those sentences. Thus, its typical perfective analysis is contradicted in those positions too.

2.3.4. As conjunctive participle

40So far I have described the behaviour of the morpheme -i in the context of grammatical auxiliaries. However, it conjoins lexical verbs too. Consider its occurrence in the following sentence:

  • 7 The conjunctive participle is taken as one of the areal and typological features of South-Asian lan (...)

41In (32) the morpheme -i conjoins two lexical verbs, bha:ng (break) and kor (scrape). In fact, it does not merely conjoin the lexical verbs but the vPs (the clauses) they project. It conjoins clauses only if they share the same subject (but not necessarily the object). Morphemes with such functions in other Indian languages are traditionally labelled as ‘conjunctive participle.’7 As for the function of the morpheme -i in (32), it seems to mark the sequence of events in time. That is, the subordinate verb (break) denotes a preceding or completed (perfected) event with respect to the event denoted by the higher verb, (scrape). However, consider the function of -i in (33):

42Notice, the conjunctive participle -i iteratively appears between the verbs occurring in a series. In (33), the first instance of -i indicates the completion of the event denoted by the verb paḍ (fall) with respect to the subsequent event a:s (come). However, the second instance of it does not indicate the temporal sequence of events, for the verb gaḍ (roll), which is marked with -i, actually indicates the manner in which the event a:s (come) happened.

43Semantically, the term ‘conjunctive participle’ is just an imprecise descriptive label for a host of functions. It can indicate succession of events, or it can indicate the completion (perfection) of an event with relation to another, or it can indicate the manner in which the event progresses, etc. Dwarikesh (1971), Masica (1976: 113-114) and (1991: 399) summarize the different functions of the so called ‘conjunctive participle’.

44Thus, there are a number of facts which show that the morpheme -i cannot be simply taken as the perfective marker for it can occur in different positions in the verb phrase with different functions. So far as the morpheme -i is concerned, the real challenge is to decide whether there are homophonous -i’s with independent and unrelated grammatical functions, or if it has a unified underlying function, and its different functions are its conditional variants. I will propose the latter perspective in the section below:

2.4. A unifying analysis of -i

45So far, it must be obvious that the morpheme -i does not have any single semantic sense. Thus, its function seems to be more general and syntactic. Descriptively, it would be adequate to say that it is a default verb subordinator, which overtly marks the subordinate relation of the lower verb with the higher one. It is marked on a lower verb to simply indicate that the lower verb needs some more verbal feature to be merged. This analysis of -i is consistent with its behaviour in all positions as it always occurs between two verbal elements.

46However, towards a more explanatory account, I would propose that its verb subordinator role follows from the more general principle of ‘distinctness condition’ proposed in Richards (2010).

47Richards (2010) proposes an obligatory ‘distinctness condition’ on multiple items of the same category label. Items with the same category label are not distinct unless their hierarchical grammatical relation is visible to Phonetic Form (PF). Therefore, they must be linearized just before the spell-out with some hierarchical grammatical information.

48Adopting Richards’ (2010) general hypothesis, I would propose that the morpheme -i in Oḍia fulfils exclusively the linearization requirement of multiple verbs adjacent to each other. So, I would refer to it as the verb linearizer (vl). It is a default grammatical connector of verbs when there is no other grammatical feature to intervene between them. Its insertion seems to be a last resort strategy to maintain the distinctness of verbs. It is inserted just before the spell-out.

49Having put the facts related to the Oḍia conjunctive participle in order, I would propose the following tree structure for (32) showing the merged structure of vP1 and vP2, the clausal projections of the conjoined verbs:

  • 8 As mentioned earlier, the object is not necessarily shared by the conjoined vPs; however, here I ke (...)
  • 9 As the movement to a theta position is prohibited, the lower subject cannot be a trace.
  • 10 In Oḍia causativized Vs can also be conjoined; in such cases, the lower V first incorporates the ca (...)
  • 11 Following Chomsky (2001) I assume that the derivation goes phase-wise; and the vP being a phase req (...)

50In (34), the two vPs are conjoined (merged) for they share the same subject. The subject sharing is expressed with the subject of the vP1 copied at the subject position of the left-adjoined vP2; the same is assumed for the shared object too.8 With the referring subject copied at the subject position of the adjoined vP2, its initial position in the vP1 is a controlled PRO;9 cf. Jayaseelan (2004: 74-75) for similar assumptions. As for the Vs, the V head of vP1 incorporates (adjoins to the left of) the V head of the adjoined vP2.10 As all the elements from the vP1 move away there remains no phonetic content to be pronounced, and the vP1 as a phase is spelt out unpronounced.11 Now, in the vP2 phase, the incorporated V-heads, namely, bha:ng (break) and kor (scrape), move higher up to incorporate the I-head (containing tense and agreement which get suffixed to the right V head); the subject moves (for Case) to the spec of IP; and in SOV languages the object somehow moves to a position higher than the tensed V. Now, at the final phase, the derivation has two V heads with the same category label, but without any grammatical element intervening between them. Here, the linearization of them, à la Richards (2010), is required as the last resort strategy; and the verb sequence is inserted with the verb-linearizer -i, just before the spell out.

51If we analyse the morpheme -i as a mere verb linearizer, then the task in hand is to explain its different aspectual functions noted in descriptive studies. I will propose that the different aspectual functions ascribed to it in the descriptive studies follow from the inherent aspectual nature of the verbs and the sequences they form.

2.4.1. Sequence of event and state: the perfective relation

52Taking the cues given in the previous subsection let us consider the anomalous behaviour of -i in the typical viewpoint aspect position in the following sentences:

  • 12 The copular auxiliaries, which bear [±Realis] world features in Oḍia, bear functions typically ascr (...)

53Oḍia has a multi-copula system. The copula -ach bears the realis feature (i.e. only present tense); and in contrast, the copula -th(a:) bears the irrealis features (other tense and mood features);cf. Mahapatra (2002)for details.12 In the analysis here the copula is [−Dynamic]. When a [+Dynamic] verb is governed by [−Dynamic] copular auxiliary, their relation indicates that a dynamic event is followed by a [±Realis] state, or a dynamic event led to a [±Realis] state (denoted by the copula). The relation in fact denotes a resultant state, and that is what is essentially denoted by the perfection of an event. This would explain the aspectual meaning of (35).

54In Oḍia, only dynamic verbs, whether telic or atelic, can be governed by the dynamic aspectual (i.e. progressive) morpheme -u. When the verb is governed by the dynamic aspectual morpheme -u, as in (17) and (19), the morpheme denotes only its dynamic progression part. If the dynamic progression is not indicated, the morpheme -u would be absent. Its absence would simply let the higher copular auxiliary be the governor of the main verb creating a non-distinct V-V sequence, forcing the verb linearizer -i to intervene, as in (18) and (20).

55Again, if a [−Dynamic] stative verb, for example so (sleep) in (36), is subordinate to the [−Dynamic] copula then there is no [+Dynamic] feature to be marked. So, there is no possibility for the progressive -u to be projected. Thus, the copular auxiliary becomes the governor of the main verb creating the same V-V sequence which calls for the intervention of the verb linearizer. Semantically, the relation between the stative verb and the copular auxiliary indicates that the state denoted by the main verb holds either in a realis or irrealis world. This would explain the aspectual meaning of (36).

56In the cases of non-copular auxiliaries, such as situation type or modal or other aspectual auxiliaries, the aspectual or the modal feature rests with the auxiliaries; so, the morpheme -i has no specific function except of a default verb linearizer (vl).

2.4.2. Sequence of events and perfective relation

57Let us come to the sequence of lexical verbs conjoined by the morpheme -i —the case of the ‘conjunctive participle’ in (32). I propose that, as a conjoiner of lexical verbs, the -i does not bear different functions, often attributed to it. Instead, these functions are necessary fall outs when two events are juxtaposed.

58If two events are juxtaposed (to share a single subject) they can be interpreted only in two possible ways: they happen either in a temporal sequence or they are concomitant. If they are interpreted to be in a temporal sequence, as they are in (32), the very nature of temporal sequence presupposes that their event boundaries cannot temporally overlap but have to be discrete; and in a sequence of temporally discrete events, an event has to be a completed (perfected) event with relation to the following one. Thus, the default verb linearizer -i which is just syntactically present in a verb sequence seems to mark the sequential function and the completion of an event with relation to the other. In fact, these two functions are the two sides of the same coin —the event sequence.

2.4.3. Concomitance of events and manner relation

59On the other hand, if two events are interpreted to be concomitant they necessarily temporally overlap. With two temporally overlapping situations one event must have the manner function for the other. The manner function of a verb can be grammatically made explicit by the verb being reduplicated in Oḍia. Consider (37)-(38) below:

60In (37) gaḍ (roll) has the manner function for a:s (come). In (38) the manner function of the verb gaḍ (roll) is grammatically made explicit by the verb being reduplicated. If two events x and y are concomitant the necessary condition is that x continues when y happens. Therefore, in (38), the x (roll) event is expected to be marked for continuity while the y (come) event happens. In fact, in (38) the reduplication of the verb gaḍ (roll) denotes the temporal extension (continuity) of the rolling event. The overt continuity marked on a juxtaposed event blocks its sequential reading with the other; thus, it indicates their concomitance.

61The other condition which they seem to fulfil is that x (roll) continues till the termination of y (come). Therefore, y (come) can take telic aspectual features like past tense, perfective, completive etc.; but x (roll) can never take them; so, it can have only the reduplicated form indicating its continuity till the end of y (come). These features ensure that the events are not in a sequence but they are concomitant and coterminous.

62Like in sequential contexts, in concomitant contexts too, such as in (37)-(38), the morpheme -i connects two verbs; and it seems to bear the manner function. But, in fact, the morpheme -i is nothing more than a default verb linearizer (vl).

2.5. Viewpoints and other functional categories

63On the basis of the discussion so far, the hierarchical position of viewpoint aspects with respect to other functional features can be represented as in (40). The structure (40) represents the sentence (5) repeated as (39):

64Like in the derivation of (34), in (40) too the ‘distinctness condition’ applies to all non-distinct adajacent Vs just before the spell out.

2.6. The perfective and its modes

65As explained with (35), the perfection of an event is expressed as the syntactic relation between a [+Dynamic] verb and the [+Stative] copular auxiliary. However, it can be noticed in (14) that the completive auxiliary sa:r (finish), which marks telicity, can optionally intervene between them. Furthermore, when the completive occurs, the copular auxiliary can be optional, as illustrated below:

66As for the completive and the copular auxiliary, both can co-occur, (e.g. 14); or, they can be mutually optional, (e.g. (35), (41)); but both cannot be dropped (in a perfective construction). Thus, syntactically, the perfective meaning is constituted as follows:

67With all the features of (42) being present, the perfective construction conveys that a dynamic event attained its telos and then gave rise to a state. Given the mutual optionality of the last two features, the perfective construction conveys different meanings. With the absence of the stative feature, as in (41), the completive simply denotes the telos of the event, but with the absence of the completive, as in (35), the stative copular auxiliary highlights the resulted state.

68The perfective features (in Oḍia) are essentially the projections of accomplishments; thus, they are grammatical indicators of accomplishments.

2.7. Dynamic (progressive) vs Stative (imperfective)

69Oḍia does not have an imperfective marker. However, Hindi, which has an imperfective marker distinct from its progressive, shows that the progressive goes with [+Dynamic] predicates but not with stative ones. The fact is illustrated below:

  • 13 Krifka, Pelletier et al. (1995) propose two types of characteristic (stative) sentences, namely ‘le (...)

70Unlike the progressive, the imperfective in Hindi goes with (characteristic) states, as illustrated below:13

71Thus, the progressive grammatically bears the [+Dynamic] feature; and in contrast, the imperfective bears [+Stative]/[−Dynamic] feature.

72The progressive in Oḍia —also in other Indian languages such as Hindi and Marathi— obligatorily selects copular auxiliaries that bear [±realis] world feature and tense feature. Without the world-time features the progressive remains uninterpretable. Consider the following Oḍia examples:

73Unlike the progressive, the imperfective does not select the world-time features. Marathi, which uses a single marker both as imperfective and progressive, construes its progressive meaning by selecting the copular auxiliary (that bears the [±realis] world feature and tense); but it construes its imperfective meaning by dropping it, as illustrated below:

74As (43), (47) and (49) illustrate, the progressive can even be saturated with the strongest temporal deictic feature, the speech time. In contrast, the imperfective in (45), (46) and (50) never accepts the speech time.Now, let us consider the Oḍia and English counterparts of (50), given in (51):

  • 14 Giorgio and Pianesi (1997: 71) assume that the AGR feature in English essentially carries tense fea (...)
  • 15 The imperfective constructions such as (45), (46) and (50) are semantically understood to be charac (...)

75Both Oḍia and English which do not have an overt imperfective marker convey its meaning by dropping both copular auxiliary and tense features in (51).14 Here, one can propose a parameter (alternatives) for the imperfective construction as follows: if a language does not have an overt imperfective marker then it can express the same by the verb not checking for the world and time features. Thus, the simple present construction of Oḍia and English and the imperfective construction of Marathi and Hindi can converge on the fact that they do not check world-time features.15

76Below, let us consider (52), the French counterpart of Marathi, Oḍia and English imperfective constructions given in (50) and (51):

77French does not have a progressive form, and the imperfective form in (52) is ambiguously used for both. The form can have progressive interpretation with the typical speech time adverb là (right now and here) appearing in the CP position; or it can have the imperfective interpretation without the speech time adverb.

78To sum up, languages (such as Marathi and French) which have imperfective markers derive its corresponding progressive meaning by selecting world time features. Alternatively, (Oḍia and English) which have a progressive marker get its contrastive imperfective meaning by dropping it and its associated world-time features.

3. Interaction between situation-types and viewpoints

3.1. Overt situation-types and the [+Dynamic] -u as frequentative

79In the previous section we considered the behaviour of the [+Dynamic] progressive when the verb projects it without the situation-type auxiliaries. However, the verb can project it with situation type auxiliaries too. The construction emphasises the situation-type, and the [+Dynamic] -u gets iterative/frequentative meaning. The frequentative reading implies a regular or irregular iteration of episodic (discrete) events; and the predicate never implies a characteristic property of the subject. Consider the following sentences:

  • 16 Let us recall here that the progressive in (21), as in (54), has a frequentative reading too. In bo (...)

80(53), in contrast with (54), illustrates that the [+Dynamic] -u cannot be projected with an overt situation type unless the main predicate is stage-level and modifiable by overt temporal adverbs or when-clauses. The iterative/frequentative reading of -u in (54) can be explained if we consider the nature of the situation type auxiliaries. When a situation type auxiliary denotes an atelic chunk of time the u cannot be bound by a single (deictic) point of time, e.g. the speech time; therefore, it needs to have an atelic iterative/frequentative reading.16 The similar reading of -u obtains with the auxiliaries la:g (activity) and ca:l (process) too. As for the accomplishment auxiliary sa:r, it always indicates a telic closure of the situation; so, the [+Dynamic] -u, which needs to express the progression of the situation, is inconsistent with it.

3.2. Role of viewpoints: Smith’s (1997) proposal

81Smith (1997: 62) proposes that: “all sentences have a viewpoint, since situation type information is not visible without one.” However, such a condition is not applicable to Oḍia. As the situation type information can be made visible with the situation type auxiliaries, sentences can occur without the viewpoint aspects. Thus, the obligatoriness of the viewpoint aspect that Smith (1997) insists on is not honoured in Oḍia. So, the following sentences are fully acceptable without any “viewpoint” morpheme:

82As summarized in §2.2, the viewpoint morpheme -u bears the [+Dynamic] feature. Thus, it holds ambiguity for all the situation types to which [+Dynamic] is applicable. With its inherent ambiguity it is incapable of making any particular situation type visible. It indicates the [+Dynamic] situation range, but it does not pick out any specific situation from that range. However, in order to make the exact situation type visible the language uses the situation type auxiliaries. The situation type auxiliaries are not obligatory in the language; however, they are selected to make the exact situation type of the main verb explicit and unambiguous. Once the exact situation type feature is given by the corresponding auxiliaries, the gross feature [+Dynamic] indicated by -u is syntactically redundant.

4. Parameter of aspect

83The aspectual values of the situation types are universal. However, languages differ as to whether they are projected in grammar morphologically or periphrastically (syntactically). Furthermore, the distinctive aspectual values of the grammatical elements differ (though in a very restricted way) from language to language. Therefore, the parametric differences are tuned to the grammatically marked aspect features.

4.1. Progressive as [+Dynamic] in Oḍia

  • 17 Kratzer (1995) suggests that individual-level predicates do not accept spatio-temporal modifiers bu (...)

84We generalized that the progressive and the imperfective contrast on [±Tense] feature. Thus, they grammatically trigger [±Tense] which is a root aspectual distinction. Further, we can generalize that the imperfective (which is [−Tense]) grammatically indicates the aspectual meaning of individual-level predicates; and the progressive (which is [+Tense]) grammatically conveys the aspectual meaning of the stage-level predicate.17 Krifka, Pelletier et al. (1995) put Individual-level predicates into two types, namely, lexical-stative and habitual-generic.

85Mahapatra (2002) shows that in Oḍia [−Dynamic] states denoted by non-verbal predicates are marked with distinct copulas. The tense-bearing stage-level copulas ach and th(a:) mark the stage-level states as well as the individual-level copula aT, which cannot bear tense and which has only the imperfective form, marks the individual-level (lexical stative) predicates. Thus, copulas too trigger [±Tense] distinction for non-verbal predicates in Oḍia.

86The progressive, i.e. the [+Dynamic], bears the root feature of all [+Dynamic] predicate types. As (1) shows, the [+Telic] feature is always associated with the [+Dynamic]; so, there is no situation which is telic but not dynamic. Thus, accomplishment, which is marked for telicity, is a special type of the [+Dynamic]. This suggests that whatever marks telicity grammatically is a sub-category of the [+Dynamic] marker. A perfective, which indicates telicity, is in fact a relation between [+Dynamic] and [+Telic]. The following tree represents the way the aspectual morphemes polarize the root aspectual features in Oḍia:

87In Oḍia, [−Telic] is a privative feature (marked with the absence of the perfective, the [+Telic] marker). Thus, without the overt presence of the [+Telic] markers (namely, completive or perfective), the [+Dynamic] morpheme becomes a sufficient feature to mark the [+Dynamic, −Telic] predicates, namely, processes and activities.

4.2. Progressive as [−Telic] in English

88We generalized that the progressive and the imperfective contrast on [±Tense] feature. So, they trigger [±Tense] as a root aspectual distinction in English too. Yet, the aspectual value of the progressive marker varies in Oḍia and English. Unlike the situation in Oḍia, the progressive in English, apart from occurring with accomplishments, processes and activities, occurs with (stage-level) states, as illustrated with the following English examples:

89Thus, the English progressive has a feature which is common to the accomplishment, process, activity and stage-level state. However, the above-mentioned situation types share neither the [+Dynamic] nor the [−Dynamic] feature. Therefore, one cannot analyse that the English progressive, like the Oḍia one, holds the dynamic feature and polarizes the aspectual features to [±Dynamic].

90We can say that the progressive in English marks the [−Telic] feature and makes the perfective mark the [+Telic]; and grammatically they trigger the binary opposition [±Telic] which seems to be a basic aspectual distinction for English. If the above observations are true for English then the aspectual values of the grammatical markers in English are as follows:

91Our analysis so far suggests that the progressive in English carries the features [+Tense, −Telic]; and the [−Tense] imperfective, which also occurs with individual-level states (which are [−Telic]) —carries the features
[−Tense, −Telic]. The [−Telic] feature being common to them they can occur with states— which are [−Telic]. However, they crucially differ with [±Tense] feature. The progressive being [+Tense] is not available with the individual-level states which are inherently marked for [−Tense] feature. Therefore, it is ruled out with the following sentences.

92As we saw, the aspectual value of the progressive marker varies in Oḍia and English. Therefore, it groups the situation type features differently and indicates different conceptual hierarchies for Oḍia and English. The basic difference between Oḍia and English seems to be that while the progressive grammatically bears the aspectual value [+Dynamic] in Oḍia, it bears the aspectual value [−Telic] in English. This possible variation in choosing the aspectual value for the progressive is a parametric choice for a language.

4.3. Root aspectual distinctions and the parameter for Oḍia

93Though it is less common for languages to show situation type grammatically, invariably languages show the viewpoint morphemes which grammatically trigger different aspectual oppositions, namely [±Tense] or [±Dynamic] or [±Telic]. This indicates the primacy of such aspect features over the situation types (Aktionsarts) in grammar. I will call the aspect features, namely, [±Tense] or [±Dynamic] or [±Telic] the root aspects and the grammatical morphemes which bear such features the root aspect morphemes. By and large, languages show that the number of such morphemes is much fewer than the number of situation types assumed. Apparently, Universal Grammar (UG) is structured using restricted binary values on which the aspectual parameter can be set.

94In fact, languages parametrically vary in assigning different values to their grammatical morphemes. The present study shows that the progressive can have either [+Dynamic] or [−Telic] as its aspectual value; therefore, it can group the Aktionsart features differently.

95A comparative difference on which languages like Oḍia can be distinguished from languages like English is the grammatical presence versus absence of the situation type auxiliaries. Predicates, in languages like Oḍia can check their Aktionsart features either with the situation type auxiliaries or with the root aspect morphemes— as redundancy (double checking) of them is avoided. Predicates in Oḍia by agreeing with the related situation type auxiliaries can grammatically check their finer aspectual features (situation types); but predicates in languages like English do not do so grammatically. Considering the fact that there are not many languages which overtly express situation type features, the absence of the situation type features in grammar is normal; however languages do not show the absence of root aspects, namely [±Tense] or [±Dynamic] or [±Telic], in grammar. Thus, their presence seems to be universal. Their universal presence seems to suggest that every language grammatically needs to check them as the most salient aspectual features of the lexical predicate; and such aspect features are grammatically available with the functional grammatical morphemes which essentially check such features of the predicates. However, languages may differ, parametrically, in grammatically setting different aspectual values for their root aspects. For instance, while the progressive in a language like Oḍia grammatically triggers [±Dynamic], in English it triggers [±Telic].

Haut de page


Brown, Roger (1973). A First Language: the Early Stages. Cambridge, MA.: Harvard University Press.

Carlson, Greg N. (1977). Reference to Kinds in English. Ph.D dissertation, University of Massachusetts, Amherst. Reproduced by Indiana University Linguistics Club, March 1978.

Chomsky, Noam (1995). The Minimalist Program. Cambridge, MA.: MIT Press.

Chomsky, Noam (2001). Derivation by phase. In Kenstowicz, M. (ed.) Ken Hale: A Life in Language: 1-52. Cambridge, Mass.: MIT Press.

Dahl, Osten (1985). Tense and Aspect Systems. New York: Basil Blackwell.

Davidson, Donald (1967). The Logical Form of Action Sentences. In Essays on Action and Events: 105-148. Oxford: Clarendon Press (2nd edition, 2001).

Dwarikesh, Dwarika Prasad S. (1971). Historical Syntax of the Conjunctive Participle Phrase in New Indo-Aryan Dialects of Madhyadeśa (Midland) of Northern India. Ph.D dissertation, Univ. of Chicago.

Giorgio, Alessandra; Pianesi, Fabio (1997). Tense and Aspect: From Semantics to Morpho-Syntax. New york: Oxford Univ. Press.

Jayaseelan K. A. (2004). Serial verb construction in Malayalam. In Dayal, V. & Mahajan, A. (eds.) Clause Structure in South-Asian Languages: 67-91. Dordrecht: Kluwer Academic Publishers.

Kratzer, Angelika (1995). Stage-level and individual-level predicate. In Carlson, G. N. & Pelletier, F. J. (eds.) The Generic Book: 125-175. Chicago and London: University of Chicago Press.

Krifka, Manfred et al. (1995). Genericity: An introduction. In Carlson G. N. & Pelletier F. J. (eds.), The Generic Book: 1-124. Chicago and London: University of Chicago Press.

Laca, Brenda (2005). Indefinites, quantifiers and pluractionals: What scope effects tell us about event pluralities. In Tasmowski, L. & Vogeleer, S. (eds.) Non-definiteness and Plurality: 191-217. Amsterdam: Benjamins.

Mahapatra, Bibhuti Bhusan (2002). Stage-level vs. Individual-level Predicates and the Four Copulas in Oḍia. Ph.D Dissertation, CIEFL, Hyderabad.

Masica, Colin P. (1976). Defining a Linguistic Area South Asia. Chicago and London: The University of Chicago Press.

Masica, Colin P. (1991). The Indo-Aryan Languages. Cambridge MA.: Cambridge Univ. Press.

Mohanty, Gopabandhu (1992). Compound Verbs in Oriya. Doctoral dissertation, University of Pune.

Nayak, Rath (1987). Non-finite Clauses in Oriya and English. Doctoral dissertation, CIEFL, Hyderabad.

Reichenbach, Hans (1947). Elements of Symbolic Logic, The Free Press, New York: Macmillan.

Richards, Norvin (2010). Uttering Trees. Cambridge, MA.: MIT Press.

Sahoo, Kalyanamalini (2001). Oriya Verb Morphology and Complex Verb Constructions, Doctoral Dissertation, University of Science and Technology, Trondheim, Norway.

Smith, Carlota. S. (1991/1997). The Parameters of Aspect. Dordrecht: Kluwer Academic Publishers.

Vendler Zeno (1967). Linguistics in Philosophy. Ithaca: Cornell University Press.

Haut de page


1 Smith (1997: 19) takes ‘static’ and ‘dynamic’ as equipollent terms; thus, their privative oppositions could be stated either as [±Static] or as [±Dynamic].

2 The state that the auxiliary rah (stay) denotes is a positional state which can be contrasted with characteristic states (typically expressed by generic sentences).

3 These Oḍia facts simply confirm Davidson’s (1967) view that an action (activity) can be seen as an event (process), but not vice versa. (9) emphasizes the action view, while (10) emphasizes its event view.

4 The VL (Verb Linearizer) -i in (16) is a default connector of verbs in Oḍia. I will come back to its nature in detail later.

5 Following Chomsky (1995) I assume that a lexically marked feature is checked (interpreted) in a condition of match/identity with some grammatical feature.

6 Sahoo (2001: 85-87) labels the -i occurring just before modal auxiliaries and light verbs as simply a “conjunctive morpheme (CM), which has very little grammatical function.”

7 The conjunctive participle is taken as one of the areal and typological features of South-Asian languages; cf. Masica (1976) and (1991).

8 As mentioned earlier, the object is not necessarily shared by the conjoined vPs; however, here I keep aside such cases.

9 As the movement to a theta position is prohibited, the lower subject cannot be a trace.

10 In Oḍia causativized Vs can also be conjoined; in such cases, the lower V first incorporates the cause feature and then incorporates the higher V.

11 Following Chomsky (2001) I assume that the derivation goes phase-wise; and the vP being a phase requires spell out.

12 The copular auxiliaries, which bear [±Realis] world features in Oḍia, bear functions typically ascribed to the reference time (R) in Reichenbach (1947: 288ff). The [+Realis] ach which selects only the (proximal) speech time has a time reference not separated from the speech time (S). On the other hand the [−Realis] th(a:), which always selects (distal) non-present time, has a time reference essentially separated from the speech time (S). Reichenbach (op. cit.) considers the reference time (R) exclusively as a temporal deictic; but he does not consider a spatial deictic, namely the reference space/world (R) referring to the distance between speech world and the event world (required for mood features). However, as natural language categories, Oḍia copulas carry Reference (R) both for time and space/world; thus, they provide evidence that R in natural languages is basically a shifter which shifts the reference point both in space and time.

13 Krifka, Pelletier et al. (1995) propose two types of characteristic (stative) sentences, namely ‘lexical stative’ and ‘habitual generic’. The lexical stative (45) has no progressive counterpart, as shown with (44); but, the habitual generic (46) does have one, i.e. (43).

14 Giorgio and Pianesi (1997: 71) assume that the AGR feature in English essentially carries tense feature too. Thus, the agreement feature (-s) in the so called “simple present” form carries present tense too. However, as the morpheme is inconsistent with the speech time (now) I do not take it as the present tense marker but simply as the agreement marker. In fact the real present (now) in English is expressed only with the support of auxiliaries, namely be or have:

(i) a. ?*John eats an apple now.

b. John is eating an apple now.

c. John has eaten an apple now.

English copular auxiliaries (namely, be and have) like those in Oḍia bear the world feature; when they occur with the present tense they restrict the event to realis present (now).

15 The imperfective constructions such as (45), (46) and (50) are semantically understood to be characteristic (stative) sentences with individual level predicates. As the imperfective does not check world-time features so do not the individual-level predicates.

16 Let us recall here that the progressive in (21), as in (54), has a frequentative reading too. In both the cases, the [+Dynamic] progressive is not allowed to be bound by a single (deictic) point of time, so it gets the frequentative reading. The same condition seems to apply to the English progressive for its frequentative meaning. See Laca (2005) for the way the semantics of frequentative is constituted.

17 Kratzer (1995) suggests that individual-level predicates do not accept spatio-temporal modifiers but stage-level predicates do.

Haut de page

Table des illustrations

Fichier image/jpeg, 56k
Fichier image/jpeg, 100k
Fichier image/jpeg, 96k
Fichier image/jpeg, 184k
Fichier image/jpeg, 132k
Fichier image/jpeg, 232k
Fichier image/jpeg, 52k
Fichier image/jpeg, 68k
Fichier image/jpeg, 144k
Fichier image/jpeg, 48k
Fichier image/jpeg, 52k
Fichier image/jpeg, 52k
Fichier image/jpeg, 104k
Fichier image/jpeg, 68k
Fichier image/jpeg, 36k
Fichier image/jpeg, 88k
Fichier image/jpeg, 92k
Fichier image/jpeg, 60k
Fichier image/jpeg, 64k
Fichier image/jpeg, 104k
Fichier image/jpeg, 72k
Fichier image/jpeg, 76k
Fichier image/jpeg, 56k
Fichier image/jpeg, 148k
Fichier image/jpeg, 44k
Fichier image/jpeg, 24k
Fichier image/jpeg, 80k
Fichier image/jpeg, 72k
Fichier image/jpeg, 64k
Fichier image/jpeg, 60k
Fichier image/jpeg, 28k
Fichier image/jpeg, 12k
Fichier image/jpeg, 128k
Fichier image/jpeg, 104k
Fichier image/jpeg, 108k
Fichier image/jpeg, 104k
Fichier image/jpeg, 96k
Fichier image/jpeg, 112k
Fichier image/jpeg, 22k
Haut de page

Pour citer cet article

Référence électronique

Bibhuti Bhusan Mahapatra, « The Parameters of Aspect for Oḍia », Recherches linguistiques de Vincennes [En ligne], 43 | 2015, mis en ligne le 01 janvier 2017, consulté le 17 mars 2018. URL : ; DOI : 10.4000/rlv.2232

Haut de page


Bibhuti Bhusan Mahapatra

Majmaah University (KSA)

Haut de page

Droits d’auteur

© Presses universitaires de Vincennes

Haut de page