2016 standard professional pro para office 365 microsoft office 2007 quicksteps ... black america 1998 serial the state and public bureaucracies a comparative
The verbal form V-ā in Hindi/Urdu: an aorist with “aoristic” meanings Annie Montaut
INALCO, SEDYL UMR8202 CNRS/INALCO/IRD) LABEX EFL
Published in Aspectuality and Temporality: Empirical and Theoretical Issues. Zlatka Guentchéva (ed.). Amsterdam: Benjamins
Although the term aorist has not been frequently used in the various grammars of Hindi, and the current label for the simple past is now “perfective”, the simple form used for the narrative past has a unique position within the global TAM system, as well as in the indicative paradigm. It displays the standard meanings usually associated to the aorist, and some of its specific meanings can help better describe this linguistic category. Distinct both from the perfective and from the perfect, it conveys the aspectual and temporal meaning of an anterior event, of eventual, gnomic present, anticipation, treated in the study within a topological and enunciative frame. The historical evolution of the verbal paradigm partially accounts for its present form and function; it also explains why it has come to convey mirative meanings, rather than the perfect.
Introduction The notion of aorist, sometimes deemed to be an aspect and sometimes a tense, has been mainly used for Indo-European languages, particularly the ancient ones. Although the term aorist has not been frequently used in the various grammars of Hindi, an Indo-Aryan language derived from Sanskrit1, and the current label for the simple past is now “perfective”, the simple form used for the narrative past has a very particular position within the global TAM system, as well as in the indicative paradigm (section 1). I will argue in this paper that this form displays the standard meanings usually associated to aorist tense in Indo-European languages, and that some of its specific meanings, similar to those displayed in 1
Although typologically distinct from it (split ergativity, head final, postpositions, serial verbs, marking only of human or specific objects, “dative subjects” for non-agentive predicates, etc. cf. Montaut 2012).
Modern Greek, can help better describe what the aorist as a linguistic category is. The Hindi simple verbal form, the only simple form in the indicative paradigm, distinct from the perfective and from the perfect, conveys the aspectual and temporal meaning of an anterior event (section 2). It also displays other meanings such as eventual, gnomic present, anticipation, i.e. a plurality of meanings which can be better studied in terms of topology (Desclés 1991) and of the theory of enunciative operations (Culioli 1991) (section 3). The historical evolution of the verbal paradigm partially accounts for its present state, both at the morphological and functional levels (section 4); it also explains why the simple form in -ā, and not the perfect, has come to convey mirative meanings, rather than inferential or hearsay meanings (section 5). The other simple forms of the indicative paradigm, similarly inherited from a (non-tensed) participle,which eventually came to convey a particular type of “aoristic effect” in its now dominant use as a counterfactual, as well as marked continuity and anti-salience in its marginal use of poetic imperfect, will be described and contrasted with the canonical aorist meanings conveyed by the -ā form in the conclusion.
1. A brief description of the aorist in Indo-European, Indo-Aryan and Hindi/Urdu 1.1. The meaning of the term and notion in the Indo-European tradition According to Meillet & Vendryes (1924), the Indo-European aorist had mainly aspectual meaning, and only secondarily temporal meaning. It represented a mere process without any consideration of duration, in contrast with the present which represented a process considered in its unfolding2. This distinction, originally strong in ancient Greek and subsequently reinforced in Modern Greek, was only present to a certain degree in Indo-Iranian. Initially, that is to say in the Greek grammatical theory of the stoic tradition which forged the term, the aorist, literally “undetermined”, is said to be devoid of any subjective value of duration or of completion (“achèvement” in Meillet & Vendryes’s translation). As a tense used for “undetermined” processes, it contrasts with “determined” tenses, which are, in this tradition, of two kinds: the tenses conveying duration (present, imperfect), and the tenses conveying completion (perfect, pluperfect). Although various meanings are classically distinguished in the uses of the aorist, such as ingressive or initial, effective, final or culminative, resultative, and gnomic,3 first in Greek, then in general, it is important to keep in mind that it contrasts basically with tenses representing both duration and completion, and that its narrative meaning of anterior event, which prevails in translations without contexts, is only secondary. In Sanskrit, this “tense” had only the latter meaning (particularly referring to recent past in ancient Sanskrit) and was mainly used in the passive voice and injunctive mood in the classical language. It never displayed the specific non-temporal meanings present in Ancient Greek and further developed in Modern Greek (Tsamadou & Vassilaki 1995). Formally, the Sanskrit aorist can be, as in Greek, marked by a -s- suffix (sigmatic aorist), or by a prefix or “augment” a- (a.bhu.t “he was” from the verbal root bhu “be”, a.bhod.it “he knew”, from the verbal root bodh “know”), but in Middle Indo-Aryan only a “radical aorist” survives, without prefix or suffix. Meillet (1908: 81) for instance quotes the Pali forms pucch.i, “he asked”, and piv.i, “he drank”, with only a person ending after the verbal root.
“Un procès pur et simple, abstraction faite de toute considération de durée” (1924: 174). For the Greek aorist Wallace (1996) adds the predictive and epistolary meanings.
The ancient or middle Indo-Aryan aorist was not retained in New Indo-Aryan, although the modern form used for an anterior event today displays a similarly reduced morphology, comparable to the Pali examples quoted by Meillet:4
The difference with the Pali aorist is that the ending directly added to the verbal basis, instead of marking a person, only marks gender and number, thus displaying a form similar to the “past” participle in Hindi/Urdu: bolā, “spoken”, in the masculine singular, bole in the plural, bolī in the feminine. In the various grammatical traditions of modern Indo-Aryan languages, since the Sanskrit aorist form was not retained in New Indo-Aryan, the category and terminology were altogether abandoned, with the exception of Nepali: the verbal form cognate to Hindi/Urdu -ā is generally called “perfect” in the English tradition (Riccardi 2003: 565), but Michailovsky (1996: note 6 p. 2 sq) also calls it an “aorist”. As for modern Persian, a language cognate to Indo-Aryan with a similar history of the form now used to refer to anterior events, this form has been treated as an aorist by Lazard (1957) and LessanPezechki (2002). Alongside this correspondence,5 there are functional arguments for considering the form V-ā as an aorist, which will be illustrated in the next sections.
1.2. The aorist in the verbal paradigm: the Hindi grammatical tradition Both Platts, author of one of the first comprehensive grammar of Urdu (1878]/1967)6 and Kellogg, author of the first comprehensive grammar of Hindi (1875/1972) name the form used, in particular, for anterior events, the “indefinite perfect”.7 The following examples illustrate its most common use for referring to an anterior event, dated or not, or a series of unchained anterior events, either in third person narration (2a) or first person discourse (2b):
stay go.AOR. MS
“He did not obtain any promotion and remained a simple gardener” (SD)
The morphology of the Modern Hindi form has a different etymology (cf. section 4), yet one cannot totally exclude a merge of both the surviving radical aorist and the past participle used for representing past events. 5 Which should of course not be overestimated, since the descriptive traditions are often loaded with cultural habits and influences not necessarily fitting the inner categories of the language studied, however it suggests a common possible reading of forms so close historically and functionally (Nepali “perfect” garyo, Persian “aorist” karde, Hindi “perfective” (initially “indefinite perfect”) kiyā (dialectal Hindi karā, karo). 6 Earlier grammars, e.g. Inshallah Khan ‘Insha’’s Daryā-e-lātāfā, uses the Arabic terminology to describe verbal categories. 7 Platts (1878: 139) reserves the term “aorist” for a form which is now called subjunctive or contingent, a simple verbal form consisting in the verb base and suffixed person endings: bol-ū̃ (speak-1sg), bol-o (speak-2), bol-e (speak-3s), whereas Kellogg (1875: 229) calls it “contingent future”.
then that.from first
“When in 1969-70 there was a break in the Congress party and Indira (Gandhi) was ousted, it gave me for the first time the motivation to…” (SD)
The reason why this -ā form is called “indefinite perfect” in the early descriptions is partly due to symmetry, since the -ā form is analogical to the perfect participle (PART2), whereas the imperfect participle (PART1) formed by the addition of -t- to the verb base before the suffixed gender/number mark, when used as a predicative form, is labelled “indefinite imperfect”. 8 It also suggests that tense is absent from their meaning, and the paradigm of the indicative system in Hindi/Urdu confirms this intuition, since tense is conveyed by the “be” auxiliary, hai for the present and thā for the past. The simple forms, when combined with present auxiliary, give respectively the “present perfect” and “present imperfect” (Kellogg and Platts’s terminology) or “present perfect” and “present” (modern terminology), and combined with a past auxiliary, the “past perfect” and “past imperfect” (Kellogg’s and Platts’s terminology) or “pluperfect” and “imperfect” (modern terminology). The progressive aspect, conveyed by a additional auxiliary (rah- “stay”) is of course available only in the imperfect system. Table 1: indicative paradigm in modern Hindi/Urdu TENSE
ASPECT Marked : V+ t“Imperfect” boltā (literary) used to speak boltā hai speak (PART1 AUX.PRES) bol rahā hai is speaking (Base AUX.PROG AUX PRES) boltā thā spoke (PART1AUX.PAST) bol rahā thā was speaking (Base AUX.PROG AUX.PAST)
Unmarked : V-Ø-“Perfect” bolā spoke [aorist] spoke bolā hai has spoken (PART2 AUX.PRES)
bolā thā had spoken (PART2 AUX.PAST)
Out of the two simple forms analogical with the participles, the “indefinite imperfect” is no longer part of the indicative system, except as a rarely mentioned variant of the general imperfect (cf. 4.3). With the subjunctive (V-Person) they represent the only simple verbal forms, apart from imperative, in the entire Hindi/Urdu verbal paradigm, a remarkable feature of Hindi/Urdu verbal conjugation, as shown in table 2.9
The only morphological distinction occurs in the feminine plural: the predicative form shows a nasal vowel, whereas the participial form shows an oral vowel: bol-ī̃ “they spoke” (feminine), vs bolī “spoken” (feminine plural and singular). The future is also written as a single word but, allowed till the 19th c., a particle could be inserted between the first morpheme (“subjunctive” karū̃), and the future auxiliary gā (“go”): karū̃ hī gā [do.1S FOCUS FUTURE] “I will sure do”, opposed to the modern form karū̃ga hī [do.FUT.1MS FOCUS]
Table 2: The global TAM system of Hindi/Urdu10 Indicative « imperfect » boltā “used to speak” (rare) boltā hai present (present imperfect) “speak” boltā thā imperfect (past imperfect) “spoke”
boltā counterfactual “would speak /have spoken” boltā hotā past counterfactual 1 bolā hotâ past counterfactual 2 “would have spoken”
bole [1s: bolū̃ ) “should/may speak” boltā ho subj. past 1 bolā ho subj past 2 “may have spoken”
Both Kellogg (1875: 224-26) and Platts (1978: 131-33) emphasize the aspectual meaning of the two participles.11 Both recognize the basic aspectual distinction pervading the whole verbal paradigm in Hindi/Urdu, since the compound tenses (in indicative, subjunctive and counterfactual moods) are derived from participles which convey aspect only. The unmarked form refers to a completed bound process which receives its temporal reference from the main verb: ublā pānī “boiled water” may refer to “water which is boiled” or “has been boiled” or “had been boiled”, depending on the context. Similarly the marked (-t) form refers to an unbound process (“in the state of”): ubaltā pānī, “boiling water”, depending on the context can mean “the water which is boiling” or “which was boiling”12. Modern terminology has not retained the labels of the early grammarians.13 The term retained for the -ā form, “perfective”, is highly controversial, since it introduces a category well represented in Slavic languages, which has little affinity with the Hindi facts (cf. section 2). Moreover, the term is now retained only for the simple form, since the compound forms are respectively labelled “past” or “present perfect” and “pluperfect”, hence veiling the unity of the three “perfect” verbal slots, and its symmetry with the three imperfect verbal slots. This symmetry is certainly less clear today than it was in the mid 19th century, where the -tā form could still be used as a general present (at least in old texts, cf. section 4). But, more importantly, Kellogg’s intuition concerning the term “indefinite” was lost. In order to grasp the form’s complete set of meanings, the notion of indefiniteness is very useful, and was intrinsically part of the classical definition of the aorist.
2. Meanings of the -ā form in narrative contexts: not a perfective, not a perfect Before illustrating the non-temporal and possibly non-assertive meanings of the simple form, let us first clarify its relation to the two other categories to which it is often equated in Indo-Aryan, the perfective and the perfect. This contrastive analysis will provide an initial interpretation of the 10
The table shows only the primary forms and excludes secondary aspects such as progressive, durative, habitual, etc., phasal Actionsart markers such as inceptive, terminative, and modalities such as presumptive. The compound forms in subjunctive and counterfactual moods differ more in aspect than in tense (the -tā hotā and -tā ho forms carry habitual meaning and refer to a state or a repeated event, in contrast with the -ā hotā and -ā ho forms, the former being moreover quite rare). As for the counterfactual simple form, it significantly conveys both present and past meanings (past and present counterfactual), whereas the unambiguous past is expressed by the compound forms. The forms of “be” auxiliary in compound forms are ho for subjunctive and hotā for the counterfactual. 11 They both consider the third participle, V-kar, presently labelled conjunctive participle or converb, as past. 12 The phonetic rule of the shwa deletion accounts for the elision of the second vowel in the form ublā (infinitive ubalnā). The French notion of “accompli” vs. “inaccompli” is convenient since it prevents any ambiguity with the perfect tense and perfective aspect, but is little used in general linguistics. 13 Although, in her first grammar, Kachru (1980) sometimes retains the old terminology of “present imperfect”, “past imperfect” and “simple perfect”. The terms used for both moods in their simple form vary without blurring the picture too much (subjunctive or contingent for the -e [-ū̃] form, contingent or conditional for the -tā form).
Mis en forme : Italien (Italie)
meaning and function of the aorist in assertions, that is to say, as a tense-aspect of the “indicative” paradigm.
2.1. Contrast with the perfective One should first note that, in languages with a clear perfective vs. imperfective opposition, this pertains not only to the perfect domain, but also to other verbal slots, particularly in the imperative and infinitive. Not only is the -ā form in Hindi/Urdu the only one labelled “perfective”, thus appearing to be the only member of the perfective paradigm whereas the non-perfective paradigm is very well endowed. Furthermore, the classical distinction, found in Slavic languages for instance, echoes a corresponding one in Hindi/Urdu, namely between simple verb and so-called “vector verbs”. In a Hindi Vv sequence, V is the main verb, conveying the basic meaning, and v is a polar (or vector) verb conveying attitudinal and aspectual specifications. The problems raised by the analysis of such polar verbs, consisting in about a dozen originally motion verbs, which are semi-auxiliarized in the sequence Vv, are far beyond the scope of the present study, and I will only summarize here the major studies of the last fifty years. In the following examples, the simple verb (in all TAM) contrasts with the compound verb Vv in the same way as simple Russian verbs contrast with prefixed verbs: for instance pishi in the imperative of the simple verb means « write » (carry on/be in the activity of writing) whereas na.pishi with perfective prefix means “write (this) down”. Similarly, delat’ “to do, act” (infinitive) contrasts with sdelat’ “execute, accomplish (do a specific thing)”. The Hindi/Urdu imperative (3) contrasts in the same way between the simple verb “do” (be in the activity of) and the sequence Vv “do+take/+give” (do a specific thing):
“You should work”
b. tum merā
do give. IMPER
do take. IMPER
“Do my work quickly!”
“Do your (own) work!”
nakonets reshilas’ deistvovat’
wrote (PERF) a.letter
The vector verb has been analyzed as a perfective marker by quite a few scholars, from Porizka (1968, 1982: perfective), who also compares it to the English particles in “drink up, drink out, write down”, to Hook (1974: relative completion) and Nespital (1997: perfective) although it is also used to specify predicate orientation (transitive verbs acquire external orientation with the vector “give”, a self benefactive meaning or internal orientation with the vector “take”, as shown by (3b)), and to qualify
the manner in which the process is performed seen from the speaker’s viewpoint (violence, impulsivity with vectors meaning “sit”, “get up”, “fall”, “throw”, “strike”).14 Vectors are rare in yes-no questions or negative statements (5a), incompatible with phase auxiliaries (inceptive, terminative), ability auxiliaries sak, pā (“be able”, “manage to”) and progressive aspect. In contrast, the vector is mandatory with adverbial specifications involving overcoming difficulties (“finally” in (5b), “ultimately”, “with difficulty”).
“When will these people come?”
“When did they come?”
“Those who were to come yesterday finally arrived”
The Vv sequence conveys the idea of carrying out a process up to its limit (culmination, final completion), which implies crossing the right boundary of the process space, hence the “intensity” effect which sometimes affects the lexical semantics of the predicate: mār d̩ ālo! for instance on main verb mār “strike” and vector d̩ āl “throw”, means “kill” whereas māro!, without vector means “strike!”. For this reason Vv perfective phrases are used for irreversible processes (representing full completion) whatever the tense, while a simple verb, even in the aorist, may allow for non-total completion since the limit of the process is not necessarily crossed:
(make give. AOR)
“The king made a painting but it could not be achieved” (from Hook 1974)
For the same reason, Vv phrases refer to telic processes, since reaching the object in a transitive process amounts to perform the process up to its limit, and telicity makes the object specific even when unmarked, as in the following example from Page (1957):
buy. AOR. MS
“He bought milk” 14
“He bought the milk (he needed, I asked him for, etc.)”
Alongside his numerous papers on the question of vectors and perfectiveness, one can refer to Nespital’s land-marking Dictionary of Hindi Verbs (1997), which contrasts simple verbs (imperfective) with compound verbal expressions involving a vector (perfective).
Since the perfective represents a process carried out up to its limits and involves the crossing of a boundary, it means that all the t occurrences of T within the process are not equivalent, whereas the aorist does not allow for differentially qualified t occurrences. Where the initial or left boundary is concerned, only the perfective, with vector “go”, can indicate an inchoative state (ho gayā be went, “became” vs the simple form huā “was”).15 The most convincing evidence for the notion of crossing the right boundary is the notion of “relative completion” claimed by Hook (1974) to be the specific feature of Vv, in complex “until” clauses (jab tak… tab tak, when till … then till, “until”, “by the time that”) involving two processes, one of which not fully completed by the time the other is fully completed:
“By the time the vegetable cooked I had reached there.” (example from Hook)
In other terms, perfective forms echo the subjective viewpoint of a speaker comparing the degree of completion of two processes (8) and emphasizing the crossing of the initial (inchoative state: cf. note 14) or final (process fully carried out) boundary. This long comparison between Vv sequences (where v conveys, among other things, perfectiveness) and the simple past form V-ā, enhances one of the distinctive features of the simple form: it does not allow for a differential orientation or qualification of a process (the so-called “attitudinal meanings” of the vector according to Abbi 1992).16 It simply presents the process as a mere process, considered “as a whole”, with no relation to the speaker and no inner differentially qualified points or occurrences of t (such as the representation of the limit as qualitatively distinct from the inner interval).17 The perfective vector, in contrast, selects the representation of the limit reached, and we shall see that the perfect represents an interval bound to the left and open to the right.
2.2. Aorist and perfect The ordinary meaning of the simple form refers, as seen above, to an anterior event, often metaphorized as “punctual”, similar to the definite past in Romance languages:18 it contrasts both with the imperfect, which involves duration or habit, and with the perfect, which has present relevance. Whereas the simple past can be represented by a bound interval, which implies a radical disjunction from the time of utterance (and from the subjectivity of the utterer), the perfect can be represented by an interval bound only on the left (point of departure: the anterior event) and open to the right, which makes it contiguous to the time of utterance. For this reason the perfect –named the “adjacent” in Greek19– fits subjective discourse and provides space for the speaker’s viewpoint: the speaker takes stock of the process and validates it. This contiguity with the time of utterance was interpreted by
The same is true with other tenses: tum jaldī t̩hīk ho jāoge [2 soon fine be go.fut.2] “you will soon get well”, vs. tum t̩hīk hoge “you will be fine”, maī t̩hīk hū̃ “I am fine”. 16 Usne yah cit̩t̩hī likh mārā “he wrote this inconsiderate foolish/rude letter (write beat)”; tum yah samajh bait̩he ki “you stupidly (mis)understood that… (understand sit)”. 17 When a Vv sequence is in the aorist, the problem consists in combining the properties of both perfective and aorist, including the contradictory properties (either instantiating or not the crossing of the boundary). One way to solve this problem is to consider the Vv sequence as lexical, as does Nespital in his dictionary of the Hindi verbs (1997). Needless to say, further research is needed in this field. 18 Whatever its designation in the respective grammatical traditions (French “passé simple” or “aoriste” (fis “I did”), Spanish “pretérito indefinido” (hicé “I did”), Italian “passato remoto” (feci “I did”). 19 This contrast between aorist and perfect was clearly stated by Stephanos of the Stoic tradition, at the very beginning of reflexions on these grammatical categories: Stephanos calls the perfect parakeimenos which means “adjacent”, and contrasts it with the aorist which is not adjacent. 15
Benveniste (1966) as typical of subjective discourse, related to the ego parameters, in contrast with the objective tense of (hi)story and narration, unrelated to the ego. Using Desclés topological model for tense and aspect, this contrast could be represented as follows (adapted from Desclés 1991)20: aorist ---[///]T1-----To aorist ---[///]T1-----[ To note
perfect ----[---] /////-----[ To
Culioli (1999: 137) accounts for the specific dissociation from the time of utterance, a feature typical of the aorist, with[ an operation which he calls ω (disjunction) and which is responsible for the radical disjunction between the Time, in relation to the Situation (Sit1, T1) of the event represented, and the Time and Situation of utterance (Sito, To), where To is related to the subject S, unlike other t occurrences of T, hence the graph T. The aorist is thus characterized by the following operation: < S1 (T1) > ω So (T To). In the perfect there is no such disjunction operation, since it must be anchored in a situation related to the utterer (or Subject). Whereas the Hindi perfect may have experiential connotations (9), and resultant connotations, particularly when the perfect participle is added (10), the aorist cannot (11):
He has gone to America (he has the experience of living there which still persists)
He has gone to the market (he is not home, you cannot meet him now)
Le crochet ouvert correspondent à To n’indique pas une coupure séparante mais une « coupure continue », c’est-à-dire qu’il y a continuité avec le temps de l’énonciation.
Examples (9) and (10) are from Narang (1984: 74, 128).
I went through the market, saw everything, bought nothing.
In (11) the speaker relates a series of processes which he represents as “objective” experiences, having no relation to his present situation (except of course for the implicit moral of this modern adaptation of a famous aphorism by the mystic poet Kabir). The temporal location is not provided by a date (2b) or time adverbial but is implicit in the basic operation associated to the simple form, an operation of radical disjunction of the event from the speaker, and the temporal succession (first predicate relates a process anterior to the second, the second to the third) is simply inferred from the succession of several predicates in the sentence: ---[///]T3-[///]T2-[///]T1-----To. ---[///]T3-[///]T2-[///]T1----[-To
Since the aorist represents par excellence the operation of disjunction from the speaker, that is to say, from both the Subject and the Time of speech, the process is not represented as validated by the speaker. For this reason it allows, in Benveniste’s terms, “the events to speak for themselves”, a notion rephrased by de Vogüé (1995) as “auto-validation”. This property accounts for the special affinity between the aorist and an objective narration of past history or tales, unrelated to the time of utterance, whereas the perfect displays special affinity with what Benveniste has called “discourse”, related to the present time of utterance. Both perfect and aorist may occur within the same discursive sequence referring to past events in the same temporal strata. The perfect, which has current relevance, is selected for summarizing the cumulative result of past processes in (12), even though the various processes (“spoil” in the perfect, “teach”, “do”, in the aorist) refer to the same temporal strata. The first predicate in the perfect takes stock of what happened, the subsequent forms in the aorist give the detailed sequence of events, in a non-chained enumeration serving as a detailed illustration for the initial outcome previously expressed in the perfect:
kar diyā hai
“Your manoeuvres have damaged our brain, tum donõ
what you taught us, you taught in the wrong way,
REL give. AOR. MS
what you gave us, you gave at the wrong moment.” (KB)
In addition to the topological property of corresponding to a bound interval cut off from the present time of utterance, the simple past also presents the distinctive property of compactness: all bound close intervals are compact (Culioli 1999: 135). This property is metaphorically alluded to in the traditional descriptions as “punctual”, something which can be represented by a point rather than a line. The open (imperfect) or semi-open (perfect) line used for representing other intervals allows for fragmentation, whereas the bound interval rules out any segmentation into separate occurrences. This does not mean that the process referred to has no duration, as is often suggested, but that no inner fragmentation is possible inside the interval disjoined from the time of utterance. Hence its peculiar affinity with historical narration as mentioned above, whatever the “real” duration of the events represented:
“King Harsha reigned for forty years.”
Example (13) shows that the aorist may also refer to a remote past, a property sometimes associated with pluperfect when it does not refer to a process which is past in relation to another past process.
1.3. Aorist and pluperfect The Hindi pluperfect can refer to a recent past as can the aorist (14), as well as to a non-recent, nonrelative past time (15):
“I just told you that…”
itne patle ho
“She[honorific] told [me] ‘you look so thin, I don’t look like your heroine,
put on some weight aur
put on some weight’, and since then I started (lit. had started) drinking beer.” (SK)
Obviously the starting of beer drinking, in the pluperfect, is not anterior to the first process in the aorist, when the renowned actress tells the beginner actor Shami Kapur that he should put on some weight in order to look like a proper hero. But the marking event which is emphasized by the actor is the advice from the actress, not the subsequent beer drinking. And when, in the same interview, the actor explains the huge success of his film “Jangli” by the fact that it was a colour picture shot outside in Kashmir, he uses the aorist, and the pluperfect for another colour film also shot outside a few months before, but not in Kashmir, the most famous place for romance; the aorist singles out a unique landmarking event compared with the pluperfect, even if the pluperfect relates to an earlier movie.
Phir usmẽ extensively Kaśmīr jāke hamne śūt̩ kiyā, jo hotā nahı̄̃ thā usse pahle, pahlī bār skrīn pe ek colour picture –colour
“Then we went to Kashmir and shot (AOR) extensively, which was not so before, for the first time there was a colour picture
on screen –colour pictures were made (imperfect) before, there were mythological or historical pictures (but filmed in a
studio). The background of ‘Tumsa nahi dekha’, it was Assam, and we did/had done (PPFT) the shooting in the local
mountains, we went there and did/had done (PPFT) the shooting. But the shooting of ‘Jangli’ was (AOR) in Kashmir, was (AOR)
in Shimla”. (SK)
Such properties already characterize the simple form -ā as an aorist, and further evidence is provided by non-narrative contexts.
3. Non-past non-narrative meanings The meanings displayed by the simple form when not referring to the past best reveal its specifically aoristic behaviour, and are further arguments for labelling the form as an aorist. As intuited by Kellogg, the form is not specified for tense (hence the label “indefinite”) but only for aspect –a feature which I will discuss later (section 5).
3.1. Aoristic meanings in non-dependent clauses: gnomic, anticipatory, effective The expression “aoristic meanings” was forged by Culioli (1999 ) for referring to a number of specific meanings of the aorist verbal form in languages which display the formal category, but which may also be conveyed by non-dedicated tense-aspect forms in various languages.22 Among these meanings, the most recurrent across languages are the gnomic present, the future of anticipation and terminatives of the type eureka “got it, understood”. Hindi/Urdu aorist can be used with general (gnomic) present reference in order to state inherent properties for instance in (17), valid for any situation (here “be a poet”):23
A — śair
poetry say.FUT.M. PL?
B — kyõ
A — “Will he (honorific) recite a poem? B — Why not, he is a poet, isn’t he? (lit. he who remained a poet)”
We also find ample evidence of the classical “anticipative aorist” described for Persian by LessanPezechki (2002) and for Greek by Tsamadou & Vassilaki (1995): for example, a bearer when answering a customer’s call, uses the aorist for signalling his immediate coming, even if he has not yet started moving towards the customer. In this situation, while French uses the present, English the present progressive, Hindi/Urdu usually uses the aorist (18a). Such meanings can occur because of the radical disjunction from the time of utterance and the absence of any link to the speaker’s actual situation, thus allowing him to represent the process not as a process in a well-actualized predication with its necessary anchoring in a situation, but as the fictive result of an already performed (compact) process.24 This also accounts for the opposite effect in the present relevance of a process such as “understand”, “find” of the type eureka, were a present perfect (as eureka itself) or present would be expected in unmarked statements.
Culioli quotes the following French examples: “Un pas de plus et je te frappe”, “Tiens ! Un oiseau”, “Voilà qu’il entre”, “Je reviendrai”, “Il faisait un pas de plus, je le frappais” (present, no verb, future, imperative, imperfect). 23 This meaning seems to be mostly restricted to the verb t̩haharnā “to stay”, but not always in relative clauses as in (16): āp amīr t̩hahare, ham garīb t̩hahare (lit. you stayed rich, we stayed poor) “you are (by destiny) rich, we are (by destiny) poor”, a situation in no way referring to the past but rather to the eternal injustice of life, prevailing for ever, from past to present and future. 24 The Grecian stoic Stephanos commented that the future meanings resulted from the same properties as the “historical” past meanings, since the aorist (lit. indefinite) does not define a given segment of time, in contrast with the perfect (adjacent) which defines a segment of time.
I am coming (right now) (Fr. “j’arrive”)
“Now I understand (got it)! I remember!” (Fr. “ça y est, j’y suis”)
These statements do not represent an anterior process constructed as a properly located predication, but a flash of awareness denoting the sudden recognition of a state of affairs. The exclamatory force usually associated with these marked statements will be further commented in relation with mirative meanings (5.1). But the aorist may also occur with no exclamatory intonation as a bare statement, with effective or ingressive meaning. (18b) is a recent title of an article mainly in the future about negotiations over the proposed new state of “Telengana”, on the first day of the meeting of the Congress Committee, and (19b) is the standard formula used when asking for the bill. (18) b. ab
“The countdown of Telengana has just started” (19) b. kitnā
“How much is it?
— All in all it’s fifty rupees”
Similarly the spectators of a big quarrel between two partners describe the end of the conflict (total renewed agreement of both partners in the same voice) in the aorist, followed by an explanation in the general present:
“Here they are both, re-united! At the first opportunity they unite against others”
3.2. Aorist in dependent clauses Yet it is in dependent clauses that the non-past, non-assertive meaning of the simple form are best observable, since the temporal reference is given by the main clause, generally in the future. It is particularly frequent in conditionals, in (21a) with a main clause in the imperative with future reference, in (21b), with the first correlative omitted, only the second element of the correlative providing for the general meaning of the sentence, in the reverse order of (21c) or in incomplete sentences as in (21d):
“If s/he falls asleep let him/her sleep” b. bāriś
“If the rain comes inside, everything will get wet”
“I will fall down if I lean on it” d. barsāt
“And in case it rains? You will slip and fall? And in case you get cold?”
In no way can (21b) be read as (literally) meaning “The rain entered then everything will be drenched” nor (21c) as (literally) meaning “I will fall so I took the help of the stick”: the apodosis initial to “then” requires the hypothetic reading, here with the eventuality meaning being conveyed by the main clause tense. For similar reasons, the aorist is the privileged form used in “beware” clauses, which in Hindi are constructed with the relative/conditional pronoun jo “who/if”:
“Beware, if you call me an unbeliever!/Don’t ever dare call me an unbeliever!”
The use of the aorist, common in these contexts (Armenian, Georgian, Greek, etc.), for expressing the category of eventual, is in conformity with its basic properties, while the future reference results here, in contrast with the “anticipative” aorist of (18a), from its relation to the main verb. One could even consider that eventuality is the core value of the aorist in dependent clauses,25 which amounts to considering it more as a mood than as tense or aspect. It also occurs, in bi-clausal sentences involving a concessive eventuality in (23) or indefinite temporality with the scanning operator bhī “ever” in (24), that the aorist constructs a predicative relation which is not stabilized (not validated) and has to take its referential value (including temporal reference) from the linked clause:
“Meanings [of these broken memories] may well (come to) emerge, still you will come to doubt whether…” (GS)
wherever appear.AOR.FS this
“Wherever she may appear, she is [like] a mirage, [if] she is on the bench, then…” (GS)
The aorist can also occur with the general clause linker ki “that”, in a bi-clausal system involving quasi-sequence and distorted logical entailment with the main verb in the future: in (25) snow will fall even before the hearer or ordinary man expects it (in December), and the aorist marks the future coming of December as a fictitious landmark, expectable yet deceptive:
“December won’t yet begin (it won’t even be December) that snow will start falling”
Contrary to the subjunctive, counterfactual and to the indicative, also allowed in the H/U hypothetic system, where they tend to occur in both clauses (always the case for counterfactuals), the aorist only occurs in the dependent clause. The fact that the main clause (apodosis), which is located in relation to the protasis, prohibits the use of aorist (preferred tense: indicative future, or imperative, or prescriptive infinitive) is one more evidence for the aorist’s basic property (not being localized). The affinities between concession (23 below) and iteration (24a) or indefiniteness have been noticed by many, and interpreted by Culioli as similarly involving scanning.
3.3. In elliptic sequences in discourse Meanings involving both non-specificity and eventuality in non-dependent clauses seem to be limited to oral contexts, often associated to lose syntax and incomplete sentences. In the following example, the Chief Minister of Delhi Shila Dikshit is asked what she usually does for relaxation with such an exhausting job. She first answers in the general present that relaxation is provided by the constant change in activities and problems, then shifts to the immediate present then to the aorist for illustrating her point:
struggle PROG PRES.1P
ho-jātā hai (…)
hotā rahtā hai
“At this moment we are struggling (PRES PROG) with water problems, then the next moment something happens (AOR) with education, then the third moment there is (AOR) some accident somewhere, and we also have to prepare 26 (obligative modality PRES) the policies, the variety which is there, that is the relaxation (…) change is ongoing” (SD)
One may reconstruct the sequence with more formal syntax, starting with an introductory “imagine that”, or “when you are a minister”, yet the fact that the present and aorist processes are on the same level (same discourse segment) remains significant: the aorist is used for processes with no duration in contrast to the present, denoting a single event which may occur or regularly occurs in a given context and then is mentioned as a strong eventuality. In a habitual imperfect context, the occurrence of an aorist with the same temporal reference amounts to pointing out a very significant fact or conclusion, for instance in the same interview when Shila 26
Similarly, the interviewer, when asking the question, suggests a comparison with ordinary people whom he represents in their various practices for relaxation by means of both the present and the aorist: hamāre jaise jo ham log ghūm rahe haĩ, jaise vo relax karne ke lie koī anulom vilom kar rahā hai Bābā Rāmdev kā, yā koī kar rahā hai ki music sun liyā, i-pod lagāke Lodi Garden mẽ ghūm rahe haĩ…: “like we people who walk (PRES PROG), like for relaxation someone does (PRES PROG) the Baba Ramdev alternate breathing, or one listens (AOR) to music, taking an i-pod we walk (PRES PROG) in the Lodi Gardens…”
Dikshit relates moments of her childhood when she and her friends enjoyed stealing small things to play hero:
“We used to rob, it was great fun”
The aorist in (27) may be read as an intensive (cf. section 5), yet in a past narrative context as above, where present reference is ruled out, it simply marks the state as conclusive and qualifying a whole series of forbidden games.
4. The historical origin of the form within the TAM system 4.1. Morpho-syntactic evolution It is a well-known fact that the V-ā form in Hindi comes from the Sanskrit past passive participle or verbal adjective -(a)ta, with regular loss of the intervocalic -t-: for instance Hindi kiyā (Sanskrit kr̥ta) “done”, from the verbal base kr̥ /kar “do/make”. The corresponding voiced consonant is still found in the modern Persian kard-, with the same origin (the Avestic past passive participle kardam corresponds to Sk kr̥ tam in the neutral singular).27 This explains why the morphology now displays the peculiarity of adding to the verbal base only gender and number endings (different of course in H/U and Sk) and no person markers, in conformity with its initial adjectival nature. The use of a passive participle or verbal adjective as a predicate explains why the perfect system in Modern H/U displays ergative alignment, with the agent in the oblique case and the predicate agreeing with the patient.28 Its origin is the well known Sanskrit nominal construction mayā tat kr̥ tam [byme(INST) this done] “I have done this”, which rapidly became generalized for transitive past processes, as was also the case in old Persian (with a genitive rather than instrumental agent: mana tyā kardam “of-me this done”). This construction was so widely generalized in Late Sanskrit that, in the 10th century Vampire Stories (Vetāla), Bloch (1906: 60) counted 1115 such expressions of past, as compared to only 35 finite verb forms, which had practically disappeared in Middle Indo-Aryan (Bloch 1906: 47-48). In this way, the expression of a transitive process referring to an anterior event came to be expressed with a predicate clearly reminiscent of verbal adjective morphology, whereas the agent, still in the oblique case, is now over-marked with a postposition distinct from the instrumental in most IndoAryan languages. Hindi/Urdu is the language which most clearly inherited this classical (pre-)ergative alignment. Other Indo-Iranian languages lost the alignment by reshaping sentences to fit a nominative alignment, some relatively early (Persian first, Bengali and Eastern Hindi in the 14-15th c.) and some more recently (Nepali). Significantly, the languages which changed alignment reshaped both the The vocalic r̥ regularly shifted to the -i- vowel, then -y- in H/U because of sandhi between radical and final vowels. Karnā is one of the very few H/U irregular verbs , and in many dialects it has been re-constructed as a regular form by analogy with all other verbs and displays the form kar-a (kar-au//i/e for the aorist and the participle 2. This is also the case in Nepali (gar+endings), with voicing of the initial consonant. 28 A construction characteristic of all Western Indo-Aryan languages and dialects, lost in the East (Bengali, Eastern dialects of Hindi). It pervades the entire verbal paradigm (including subjunctive and counterfactual) and is required whenever the verb takes on the form of the second participle, or simple -ā form. Exceptions are few and correspond to the classical ones found cross-linguistically. See for more details Montaut forth. 27
syntactic alignment and the predicate morphology, by adding personal endings (-am, -i, -e in Bengali for instance for persons 1, 2, 3, -ẽ, -is, -yo in Nepali, -am, -i, -e/∅ in modern Persian). Moreover, both changes occurred during the same period.29
4.2. Semantic aspects of these changes as concerns tense and aspect As the nominal sentence for past transitive processes generalized, it started to lose its “expressive” or “stylistic” meanings, which originally emphasized the result and not the process. It then acquired wider meaning, encompassing both process and result (“various nuances of past tense and modality” in Bloch’s 1906 terms). The original meaning of the passive past participle, a state, can be represented as an open space, not taking any boundary into account, as opposed to the anterior which only takes into account the bound interval (event) in disjunction from the utterance time, and to the perfect, which represents the adjacency of the resulting state with the event which produced it, allowing for the topological representation below (which can be contrasted to 1.1, also adapted from Desclés 1991), with the greyish part showing the segment mapped by the verb (open interval): state
When the participles generalized they had an “open meaning”, a notion borrowed from Bybee (1994) covering the domains of anterior event or preterit, perfect and resultant state.30 This very broad extension of the notional domain mapped by the verbal adjective is linked to the dismantlement of the verbal paradigm in MIA, where it is generally assumed that only two or three finite forms out of the some 20 Vedic forms were retained (present, imperative, and, in certain regions and stylistic registers, future). As the verbal adjective grammaticized as the unmarked expression of past rather than as a stylistically marked stative-passive, the clause became more and more perceived as an active predication since there was no other option, and lost the passive meaning and patient orientation. Later, the need was felt in certain statements to avoid ambiguity or to emphasize the resulting state, therefore a new form was created by adding a copula, originally expressive then grammaticizing in the meaning of resultant state. Initially the copula occurred in the first and second person to prevent agent ambiguity (Bloch 1906):
who.INST beat.PART2.NOM. MS
“By whom have you been beaten?” b. tenāsmi [tena.asmi]
“by whom have you/we/they/I/he been beaten?”
“I (not you, not they) have been told this by him” = “he told me”
In Bengali the suffix -l-, now read as a past marker (anterior event) as in por.l.am “I read”, was originally a nominal/adjectival suffix, the same as in rangīlā “colourful” (rang “colour”, kant̩īlā “thorny” (kan̩t̩ “thorn”). Other IA languages can be considered as transitional, such as Marathi (which has one single person ending for this tense in the second singular person, as in Panjabi, and does not mark the agent in the first and second person, although agreement is always with the patient (with patient and agent in the second person). Open meaning often corresponds to unmarkedness, which is not to be equated with zero-marked forms: the present in French is an unmarked form with open meaning (specific and general), whereas the general present in English is a zeromarked form with restricted meaning (specific, excluding the meaning “general”).
The copula later came to emphasize stativity (to prevent another kind of ambiguity, event or state) or was used to simply introduce stylistic variation according to Breunis and Bloch, and as soon as this originally stylistic variant became more expressive of state or “condition”, it ceased to behave as a stylistic variation and became a grammaticized expression of perfect or the resultant state of an event (Breunis 1990: 141). At the same time, the simple form restricted its previously “open” meaning to the expression of anteriority (event: preterit). This echoes the renewal of the present (first competing with a new progressive marker for that specific meaning, then retaining only the other meaning). The situation found in early NIA similarly shows open meaning, which was probably in the process of becoming restricted before the competing copula form, whereas the contemporary situation clearly shows strictly complementary distribution of both forms. If one agrees with Bybee (1994), one may analyse this as emergence of a zero marker with anterior meaning, whereas previously the unmarked form had unspecified meaning in the whole past system. Obviously, when the former participle is used as a predicate for representing events, even if the agent remains in the oblique case as in passive sentences, the emphasis is more on the process (source oriented) than on its result, and the whole statement is more and more perceived as active.31
4.3. The non-past domain: -e and -tā forms The only finite verbal form retained throughout MIA, apart from the imperative, is the Sanskrit present, which therefore came to occupy a wider domain since there was no competing form in the non-past (except, in certain regions, the old sigmatic future), and then to restrict its meaning. The Sanskrit indicative present was formed by suffixing person endings on the base, which were phonetically reduced as follows for the verb čal “walk”: calati (go.PRES.3MS) > calaï > calai > cale calanti (go.PRES.3MP) > calaĩ > calẽ32 This form (V.e) is now interpreted as a subjunctive (cf. table 1) with optative meaning in independent clauses and various non-assertive meanings in dependant clauses. But until the 19th century it had wider meaning, covering the whole domain of non-past, since hardly any other form was available for future, eventual, and present, whether specific or habitual.33 This new broader meaning can be regarded as open meaning (Bybee 1994), covering several restricted meanings which were later distinguished. The form ve khelẽ [play.3MP], today subjunctive, was still translated by “they play” by Kellogg in 1875 and described as a contingent which could be used as a general present. It is still found in the literature of the time with present meaning, although this became progressively rarer in the 19th century,34 and only retains this meaning today in proverbs, which are well-known for retaining archaic forms (including the older phonetic variant of the 3rd person -ai) in (29). Hence the association of the aorist meaning to this form (“perfect”) by Kellogg:
“You (will) reap what you sow”
“One person does and another one benefits”
The ergative agent, not the patient, retains the syntactic properties of the subject (control properties). Nespital (1986: 145) sees in this semantic/discursive shift the emergence of a “Neuer Proto-aktiv Satz”, already present in Pali (early Prakrit) in the Buddhist text of Milindapanha. 32 The second person, originally ending in -asi, also evolved into a final -e (-asi > -ai > -e), like the third person. The first, āmi, has a more complicated story. Analogy reshaped the modern plural forms where original endings disappeared, except the third person plural anti > aĩ > ẽ. 33 Except the Sanskrit sigmatic future, retained in certain regions only (-syati > sai/hai). 34 For instance, it occurs only rarely in the texts written by the language teachers of the Fort William, College who were supposed to set the modern grammatical standards during the first decade of the 19th century (1800-1810).
A first periphrastic form in -tā hai (lit. “is … -ing”) came to be used initially for emphasizing the actualized present in contrast to the future and contingent present, a form which maintained open meaning (progressive and general present) until the mid 19th century. The modern contrast between habitual/generic and progressive such as in Table 1 was still not fully grammaticized in Kellogg’s time, he glossed the first form by both present meanings, and the new longer periphrasis in rah hai (lit. “is stayed”) by a marked periphrastic expression (“be engaged in”): ve khelte
[3P play.PART1 be.PRES.3MP]
“they play” or “they are playing”
haĩ [3P play stay.PART2 be.PRES.3MP]
“they are engaged in playing”
This means that, in the middle of the 19th century, the -tā hai (“is …-ing”) form, today a “general present” still had its progressive meaning (in conformity with its morphology), along with general/habitual meaning.35 When the rah form lost its literal meaning and came to be required for the expression of the progressive, then no longer perceived as an expressive device, the -tā hai form became restricted in meaning to habits and generics, losing its open meaning. This process, whereby the original meaning was broadened and then restricted because a new marked form emerged to remove ambiguity, or as a stylistic/expressive device, is similar to what happened in the past domain for the simple form.36 The three simple forms of the H/U verb paradigm underwent a similar process of widening (“open” meaning) then restricting their meaning. The third, the -tā form, derived from the Sanskrit present participle -anta (> ata > at) never came to dominate in the non-past domain as did the -ā form in the past domain because of the resilience of the finite Sanskrit present. On the other hand, the emergence of the new periphrasis for the present, the -tā hai form, explains why the simple -tā form ceased to be used in the indicative in H/U,37 although it still occurred in various regional varieties in early Hindi texts such as Old Rajasthani in the West (30a) and Old Awadhi in the East (30b), with or without the addition of the new gender/number marks:
Texts from Fort William College around 1800 illustrate this situation, where the rah form is not yet fully grammaticized as a progressive, but still retained stronger emphatic and literal meaning (“engaged”), used as an optional stylistic or disambiguation device. For instance, in 1810, Lalluji (Premsāgar, 4th chapter) uses, in the same context, a -tā hai (a) form and a rahā hai (b) form for describing an obviously actual and not habitual process, with ostensive indications, when Krishna suggests to the cowherd that he visit the nearby Brahmins from Mathura, who were at that time celebrating a sacrifice (the smoke being visible): Texte publié en hindi sans date. a. dekho jo dhuā̃ dikhāī detā hai tahā̃ mathuriye Kans ke d̩ar se yagya karte haĩ (general present) “Look, the smoke which one can see, there Brahmins from Mathura are celebrating a sacrifice for fear of Kansa” b. vahā̃ gae jahā̃ mathur bait̩ he yagya kar rahe the (progressive present) “We went where the Brahmins from Mathura were celebrating a sacrifice” The present Hindi expression for Kellogg’s translation of the second form (“were engaged in playing”) would be something like ve khelne mẽ lage haĩ (they play.INF in stuck are). 36 This can be compared to the loss of open meaning of the English present: the simple, unmarked form “drinks” originally had both general and specific meanings, but when a marked form, originally optional and used as a stylistic or disambiguation device, became obligatory for expressing (restricted) progressive meaning, the older form “drinks” restricted its meaning to the general present (Garcia & Putte 1989). This process has been explained as a conventionalization of the inference which, in conformity with conversational rules, constrains the listener, in the absence of the periphrasis, to rule out the marked, more specific, meaning (Carey 1994). 37 In Sindhi however the -ta/to form is used as a future (Beames 1871: 126, Trumpp 1872). And in Garhwali with personal endings suffixed to the -ta (> da) form it is used for the general present: mi par̩hān̩o ku kām kardū̃ [1SG teach.INF of work do.PART1.1S] “I do the job of teaching children”. 35
“He takes his ritual bath in a pond.” (Chand Bardai’s Prithvirāj Rasau, 12th-13th c.) b. purus̩
“Men say” (Tulsidas’s Ramayana, 16th c.)
In modern standard Hindi/Urdu however, the -tā form is restricted to the expression of counterfactual meaning, in both clauses of conditional sentences, with present or past reference –blurring the symmetry of the indicative paradigm (31). This is in conformity with the basic meaning of the -t- form (imperfect, then not realised, then not realisable) and may be considered to have “aoristic meaning”, similar to the French imperfect in sentences analyzed by Culioli (1999) such as “Il faisait un pas de plus, et je lui cassais la figure” (“had he made one more step, I would have beaten him up”).
“If I had been/were afraid I would not (have) come here.”
The -tā form also marginally occurs in modern standard Hindi/Urdu as a short imperfect, sometimes called truncated, in the indicative, rarely mentioned and considered a habitual (cf. table 1). In a language with two other forms for the past habitual (the -tā thā or general imperfect and the auxiliated frequentative), the specificity of the “short” imperfect lies rather in transforming an event or more often a series of events into a state-like situation, which must take its temporal reference from a long imperfect in the previous context. The -tā form in such contexts qualifies the processes as typically non-salient, almost as nominal extensions of the initial process previously expressed by the long imperfect. For more details, see Montaut (2006b). This anti-salience feature makes it symmetrical with the other participial form, the aorist, which is typically used for salient processes.
5. Mirative meanings of the aorist Apart from the canonical meanings analysed in section 3 (anterior, general truth, eventual), the aorist may also convey sudden awareness, with direct reference to the time of utterance and immediate relevance for the speaker in marked statements whose unmarked counterpart would be in the present or present perfect. These special meanings include sudden changes of state, surprise and salience, polemic or irony.
5.1. Sudden change of state Examples (18a) above (samjhī “understood”) and (20) are typical of the aorist’s terminative meaning, which accounts for a definitely and markedly achieved state rather than asserting an anterior event. It can be compared with the perfect-present used by ancient Indo-European languages in such contexts. Similarly, the very common formula yah rah/ā/rah.ī X, “here is X”, does not obviously refer to an anterior event, but simply marks, in contrast with the present perfect or present, that the new situation strongly contrasts with a previous one:
“Here is your tea, it’s nice and hot, drink it!” (Fr. Voilà ton thé!)
Such statements often have exclamatory intonation, a phonetic pattern typically used for high degree, hence the “emphasis” or “intensity” intuitively associated with the aorist. For instance, when the predicative notions “remember” or “forget” are expressed in the aorist (33a), there is marked emphasis on sudden awareness of a situation, which is not the case with the present or present perfect (33b):
jay mātā kī !
glory mother of umbrella
bhūl gayā! gayā
“By god! I forgot my umbrella!” b. ab
ho gayā hū̃ ,
/ forget go
“Now I have become an old man, I have forgotten a lot/I forget a lot”.
More than marked termination, what characterizes the representation of such processes with present relevance in the aorist is this feature of sudden awareness in the speaker’s consciousness, which also accounts for the so-called “inchoative” meanings of the aorist with verbs expressing a change of state:38
In Hindi, changes of state such as in (34) may even be lexicalized in quasi-ritual expressions which significantly correspond to non-verbal formulas in French or English, such as “pleased to meet you”: “enchanté/ravi”, “sorry”: “désolé”: in Hindi khusī huī (pleasure was), afsos huā (sadness was) are practically always in the aorist, although they have direct relevance to the situation of utterance. Rather than the crossing of a boundary, what is at stake here is the (already completed) passage to a new state, since theoretically there is no complementary interval containing the boundary (Culioli 1999: 135), that is to say, no boundary can belong to the complementary linguistic interval.
(ā rahā hai)
great pleasure come.AOR.MS
“What fun! /Great fun!”
(“[I] am enjoying it”)
In such processes, which can be expressed by non-dedicated tenses in other languages (Peul for instance uses the general past for cognitive states, emotions and feelings,39) French tends to use nominal structures such as “pigé/compris” (got it), “super !/le pied” (super!/wow), what is emphasized is that the right (terminative) or the left (inchoative) boundary has already been passed and only the (markedly) new situation is represented. Such statements do not construct a stabilised predication about something, they rather present a new state as a whole, contrasting with a then particularly salient initial situation.
5.2. Surprise and salience The H/U aorist also displays more typically mirative meanings when it is used for expressing either surprise or the mere salience of some event. Statement (35) in the aorist, when it does not refer to a past event in a narrative, but has direct present reference, can only be uttered when the speaker, who remembers the girl as a small child, is suddenly confronted with an unexpectedly tall girl. The use of the perfect instead of the aorist makes the statement a simple assertion devoid of surprise:
lambī ho gayī!
(bar̩ī ho gayī hai / bar̩ ī hai)
(tall be go PFT
/ tall be.PRES)
“Wow, the girl has become so tall!” (what a change) / “How tall the girl is now!” (Fr. Qu’est-ce qu’elle a grandi ! Mais c’est qu’elle a grandi !)
Significantly, such statements in the aorist can be repeated in the perfect when surprise is ruled out, as a standard assertion, already integrated in the speaker’s conceptual framework, and the perfect makes it possible to further elaborate on the process and comment on it (36):40
A — are!
hair ripe go.AOR
B — hā̃ hā̃ ,
pak gae haĩ, par fizūl
yes yes hair ripe go
but useless NEG
A — My god, (but) your hair is white now (have got white hair now)! B — Yes yes, my hair has whitened, but not in vain (my ageing was fruitful too)!
The perfect takes into account the further consequences of the process, whereas the aorist only expresses sudden awareness of the event, and is uttered by a speaker who gives no further thought to the here and why of the event.
39 Quoted in Culioli (1999: 137) with reference to Arnott’s grammar of the Peul language. 40 This example, like (38) and (39), are from Montaut 2005 and 2006a where other examples
It is also significant that similar statements, when simply representing a salient event which occurs in a kind of blank, do not so much deliver information as comment on its salient (exciting, interesting) nature, as in (37), adapted from Culioli’s example “Tiens ! Un oiseau !” given as an illustration of the aoristic effect:41
“Hey, look here, a rabbit!”
(“A rabbit is coming out”)42
Equivalent statements involving surprise and salience, which are expressed by the evidential form in languages with an “evidential” grammatical category like Armenian (Donabedian 1996), are often expressed by other non-specific means such as a non-final accent and special word order in Russian: Donabedian & Bonnot (1995) and Bonnot (2004) have shown that the intonational pattern, which makes the statement a non-segmentable unit, echoes the construction of the statement as an already pre-constructed predication taken as a whole, contrary to ordinary assertions which need to be constructed in relation to a reference point and can be segmented into and . The fact that in H/U the aorist behaves like one of these possible means for embedding a pre-constructed statement, in a subjective utterance with exclamatory force and a meaning of surprise, is partly because it points to “absolute” events not related to any precise temporal reference nor to a subject organizing the representation: a speaker who consciously processes and organizes the representation of an event constructs its representation by posing a landmark, in relation to which the predication is localized, hence the possibility of segmentation. Not surprisingly, the H/U aorist is totally devoid of inferential or hearsay meanings, common in languages which have an evidential marker derived from the perfect, since the perfect only allows for a space adjacent to the utterance situation and this space can be used to emphasize distance from the source of information (Guentchéva 1996). Whether the feature “suddenness of awareness” is a central feature of evidentiality, as suggested in Slobin & Aksu-Koç (1982), or pertains to mirativity as a specific independent category, as suggested by DeLancey (1997), goes beyond the limits of this study. But it should be noted that it is a central feature in the Turkish evidential mis (Aksu-Koç & Slobin (1986)), and that it requires the selection of the marker red in Tibetan, used for new information which has not yet been integrated into the speaker’s conceptual frame, as opposed to yin/yod, for already wellintegrated information.
5.3. Controversial and ironic meanings An unmarked assertion in the perfect stating the present result of anterior events may be reasserted in the aorist, as a quasi-quote, obviously not with a meaning of surprise, but in order to discard the previous assertion, with hints of denegation, irony or polemical emphasis. In the following dialogue, for instance the first clause (38a) expresses surprise at a friend’s visit, in the aorist (cf. 5.2), the second one (38b), by the visitor, gives the reasons for the visit in the perfect, first with the same verb (“come”) then in a different specifying predicate (“a problem has come”) [a description of the problem follows, which is the love-marriage of his son], and the last one (38c), uttered by the first speaker, dismisses this “problem” ironically in the aorist:
Cf. note 19. Statements similar to (35) or (37) would be represented in the so-called “inferential perfect” in the closely related IA language Nepali (cf. Michailovsky 1996, who cites “The lake is (inferential.perfect) beautiful”, with the meaning “What a beautiful lake”).
“Hey what brings you here so early? b. maĩ
mere sāmne bar̩ ī-bar̩ ī samasyā
— I have come to ask you for advice, today I have encountered a very serious problem […]
to bar̩ ī khuśī
happiness of thing be.PRES that
— Why, what is the problem with that, it is on the contrary a pleasure that…” (KPB)
The re-expression in the aorist in (38c) of what the second speaker said about the problem in the perfect (38b) has no other meaning than the ironical disagreement of speaker 1, who discards his friend’s representation of the facts as catastrophic. When re-uttered by the same speaker, the aorist transforms the previous statement in the perfect into a stronger judgement, negative in example (39), perceived as more emphatic, and in fact not pertaining to the same level: the perfect asserts a (negative) state of affairs, the aorist makes it a subjective evaluation, no longer an assertion but a personal commentary:
“My god, very bad days have come. Young people say ‘Outcasts and Brahmins, all are equal’. By Ram, what rotten times!” (we are experiencing) (KPB)
The disagreement can be far more controversial and aggressive in other contexts, particularly with the quasi-lexicalized form bar̩ ā āyā (big come.AOR), used as an insult to dismiss the pretention of a fake, or one so considered by the speaker (“child of”):
(* bar̩ ī
* bar̩ ī
“Out of here you! Who are you to threaten me?”
Conclusion In narrative contexts, the Hindi/Urdu aorist, a form which inherited adjectival morphology from the Sanskrit participle, is essentially dedicated to representing anterior events. In non-narrative contexts, it also systematically represents eventuality, especially, but not only, in the protasis of conditional sentences. It is also used for other fictive representations, such as the anticipative future in nondependent clauses, and, more marginally, for the gnomic present. Its specificity in Hindi/Urdu, comparable to Greek, lies in its salient and mirative meanings allowing for representing processes or states with immediate present relevance. The basic operator triggered by the unmarked -ā form in H/U entails a radical disjunction from the time of utterance and from the subject (“énonciateur”), an operation typical of the aorist cross linguistically. This operation accounts easily for the first meanings (narrative anterior, possible, gnomic), which moreover are generally displayed by languages having an aorist.
Yet it appears first as a contradiction that the aorist may characterize processes with direct relevance to the speaker’s present, in examples such as “I have understood”, usually represented by perfectspresents and typically corresponding to the standard values of the present perfect, or as in “I am so hungry!” representing a sudden change of state. In many languages the perfect, and not the aorist, may allow for such variations in meaning. The (present) perfect may be used as an anterior in Tswana (“révolu”: Creissels 1999), and similarly in contemporary French, where the label “aoriste du discours” has been coined to account for this seemingly contradictory meaning of the perfect. Many such examples are given in the special issue of Chronos on Past and Perfect edited by Carlier, Lagae and Benninger (2000). This polysemy of a past form involving “révolu” (completed) and “accompli” (perfective) echoes the various meanings of the very term past and praetere, the etymon of “preterit”, studied by Mellet (2002), De Vogüé & Mellet (2004) and further commented by De Vogüé (2006): she brilliantly shows that this verb occurs three times in a short passage from Saint Augustine with three distinct values of attested/has-occurred (“avéré”), vanished/has-been (“disparu”) and passage, echoing the three meanings of the term and notion “past”.43 In Hindi/Urdu it is the aorist and not the perfect which appears to allow for the most extended polysemy, since it can convey the three above mentioned meanings, and the reason for this may be its former use as a perfect with open meaning (4.2). The mirative (and never inferential or hearsay) meanings, which are now attached to the aorist (and never to the present perfect) are part of an even wider polysemy, possibly related to the history of the form which was originally devoted to the representation of states, and its present morphology (a participle) is still reminiscent of this. But a more robust explanation lies in the basic operation triggered by the aorist. Its compactness makes it fit for the representation of inner kinesthesis which cannot be rationally processed and verges on unrepresentability (the unspeakable), whereas the radical disjunction from the time of utterance (“operation de rupture”) makes it available for the representation of predicative relations, which are not constructed in relation to the Subject and time of utterance like ordinary assertions, but are already given as a pre-construed whole. What varies is the range of values differently construed according to the different speech contexts, what does not vary (the “invariant” in Culioli’s theory) is not a universal value but the possibility of being part of this particular mind set (De Vogüé 2006). Interestingly, the variations of a given form (here the aorist) within a given language (here Hindi/Urdu), crosses through variations in forms dedicated to other “central” meanings in other languages, which amounts to considering intralinguistic variation as a possible way of broaching inter-linguistic variation.
References Abbi, Anvita. 1992. The Explicator Compound Verb. Indian Linguistics 53: 27-46. Aksu-Koç, Ahyan & Slobin, Dan. 1986. A Psychological account of the development and use of evidentials in Turkish. In Evidentiality: the Linguistic Coding of Epistemology. Chafe, Wallace & Nichols, Johanna (eds). Norwood, New Jersey: Ablex, 159-167. Beames J. 1871 , A Comparative Grammar of the Modern Aryan Languages of India. Delhi: Munshiram Manoharlal. Benveniste, Emile. 1966. Problèmes de linguistique générale. Paris: Gallimard. Bloch, Jules. 1906. La phrase nominale en sanscrit. Paris: Champion. Bonnot, Christine. 2004. Représentation préconstruite et focalisation: pour une analyse unitaire des énoncés à accent non-final en russe moderne. Slovo 30-31: 211-270. Breunis, Andries. 1990. The nominal sentence in Sanskrit and Middle Indo-Aryan. Leiden, Netherlands: E.J. Brill. Burton Page, John. 1957. Compound and Conjunct Verbs in Hindi. BSOAS, p. 160-178. 43
The English translation for these French notions (révolu, avéré, accompli, disparu) is so problematic that I give them in French with tentative English equivalents.
Bybee, Joan. 1994. The Grammaticalization of zero. Current Issues in Linguistic Theory (Perpectives on grammaticalization) 109: 237-254. Carey, K. 1994. The Grammaticalisation of the Perfect in Old English. In Pagliuca N. (ed.). Perspectives on Grammaticalization. Amsterdam: Benjamins, 103-116. Carlier, Anne, Lagae, Véronique, Benninger, Céline (eds). 2000. Cahiers Chronos 6 (Passé et parfait). Amsterdam: Rodopi. Creissels, Denis. 1999. Parfait et statif en tswana. Cahiers Chronos 4 (La modalité sous tous ses aspects) : 185-202. Culioli, Antoine. 1999 . Valeurs aspectuelles et opérations énonciatives: l’aoristique. In Pour une linguistique de l’énonciation 2 : Formalisation et opérations de repérage. Gap : Ophrys, 127-143. De Vogüé, Sarah. 1995. L’effet aoristique. In Bouscaren, Jacqueline, Frankel, Jean-Jacques et Robert, Stéphane (eds.). Langues et langage. Problèmes de raisonnement en linguistique. Paris : PUF, 24759. –. 2006. Invariances culioliennes. In Antoine Culioli, Un homme dans le langage. Colloque de Cerisy. Gap: Ophrys. De Vogüé, Sarah, Mellet, Sylvie, & al. 2004. Aspect : entre parole, langues et langage. In Dictionnaire européen des philosophies. Cassin, Barbara (ed.). Paris : Seuil/Le Robert, 116-144. DeLancey, Scott. 1997. Mirativity, the grammatical marking of unexpected information. Linguistic Typology 1-1, 33-52. Desclés, Jean-Pierre. 1991. La prédication opérée par les langues (ou à propos de l’interaction entre langue et perception). Langages 103 : 83-96. Donabedian, Anaïd. 1996. Pour une interprétation des diverses valeurs du médiatif en arménien occidental. In Guentchéva 1996 L’Enonciation médiatisée. Louvain : Peeters. : 87-108. Donabedian, Anaïd & Bonnot, Christine. 1995. Lorsque la morpho-syntaxe rencontre la prosodie : accent non final en russe et médiatif en arménien. Faits de Langue 13 : 182-90. Garcia E. & F. van Putte. 1989. Forms are silver, Nothing is Gold. Folia Linguistica Historica VIII-1-2: 365-384. Guentchéva, Zlatka (ed.). 1996. L’Enonciation médiatisée. Louvain : Peeters. Hook, Peter, E. 1974. The Compound Verb in Hindi. Chicago: University of Michigan Press. Kachru, Yamuna. 1980. Aspects of Hindi Grammar. Delhi: Munshiram Manoharlal. Kellogg, R. 1972 . Grammar of the Hindi Language. Delhi: Orient Reprints. Lazard, Gilbert. 1957. Grammaire du persan. Paris : Klincksieck. Lessan-Pezechki, Homa. 2002. Système verbal et deixis en persan. Paris : L’Harmattan. Meillet, Antoine. 1908. Sur l’aoriste sigmatique. In Mélanges offerts à M. F. de Saussure. Paris : Société de Linguistique de Paris, p. 81-116. Meillet, Antoine & Vendryes, Joseph. 1924. Traité de grammaire comparée des langues classiques. Paris : Champion. Lessan-Pezechki, Homa. 2002. Système verbal et deixis en persan. Paris : L’Harmattan. Mellet, Sylvie. 2000. Le parfait latin : un praeteritum perfectum. Cahiers Chronos 6 : 95-106. Michailovsky, Boyd. 1996. L’inférentiel du népalais. In Guentchéva, Zlatka (ed.). L’Enonciation médiatisée. Paris/Louvain, Peeters : 109-123. Montaut, Annie. 2005. On the evidential meanings of the Hindi aorist, Tohfa-e-dil, Festschrift Helmut Nespital, D.W. Lönne ed., Verlag fur Orientalische Fachpublikation, Reinbek, band 1, vol. 1, p. 345-364. -. 2006a. Mirative Meanings as Extensions of Aorist in Hindi/Urdu. The Yearbook of South Asian Languages and Linguistics 2006 (ed. Rajendra Singh). Berlin/New York : Mouton de Gruyter, 7186. -. 2006b. Figures du sujet : deux cas particuliers du continu et du discontinu. In Antoine Culioli, Un homme dans le langage, Colloque de Cerisy. Gap : Ophrys, 187-208. -. 2012. Le Hindi. (Société de Linguistique de Paris).Louvain: Peeters
-. Forthcoming. Why the ergative case in modal (in)transitive clauses? The historical evolution of Aspect, modality, ergative and locative in Indo-Aryan. In Stronski Krysztof & Eystein Dahl (eds.). Ergativity. John Benjamins. Narang, Vaishna. 1984. Communication of situation/time in Hindi. Delhi: Garimashri Prakashan. Nespital, Helmut. 1986. Zum Verhaltnis Genus Verbi, Nominativ- und Ergativ- Konstruktionen in Indoarischen aus synchroner und diachroner Sicht. Münchener Studien zu Sprachwissenschaft 47: 127-158. -. 1997. Hindī Kriyā Koś. Dictionary of Hindi Verbs. Allahabad. Lokhbarati. Platts, John. 1967 . A Grammar of Hindustani Language. Delhi: Munshiram Manoharlal. Porizka, Vincenz. 1968. On some Aspectual Functions of Simple Verbs and Verbal Expressions in Hindi. In Singh K.S. (ed.). Readings in Hindi-Urdu Linguistics. Delhi: National Publishing House, 158-64. -. 1982. On the Aspectual Contrast in Hindi. In Koul, Omkar N. (ed.). Topics in Hindi Linguistics, vol.2. Chandigarh-Delhi: Bahri Publications, 122-134. Riccardi, Vittore. 2003. Nepali. In Cardona, George & Dhanesh, Jain, The Indo-Aryan Languages. London/New York: Routledge, 538-80. Slobin, Dan I. & Asku-Koç Ahyan A. 1982. Tense, Aspect and Modality in the Use of Turkish Evidential. In Hopper, Paul J. (ed.). Tense-Aspect: Between Semantics and Pragmatics. Amsterdam/Philadelphia: John Benjamins, 185-200. Trumpp, Ernst. 1872. Grammar of the Sindhi Language. London-Leipzig. Tsamadou, Irini & Vassilaki, Sophie. 1995. Aspects du grec moderne. Lallies 15 : 7-69. Wallace, Daniel B. 1996. Greek Grammar beyond the basics. Grand Rapids: Zondervan Publ. written corpus: Kamleshvar Prasad Bhatt, Kaphalpān̩ī, 1990, Krishna Baldev Vaid : Theatre (200006), Geetanjali Shree : vahā̃ hāthī āte the (2010) oral corpus: interview with Shila Dikshit (minister), interview with Shami Kapur (film actor); conversations with middle class Uttar Pradesh speakers 40-50 years old, recorded in Delhi and Paris, 2008-10.