Mirative Status of the Urdu Discourse Marker to

1. INTRODUCTION

It is often observed that the form-function relation at clause level is not always one-to-one and languages may use ‘strategy’, instead of grammatical ‘category’, to express certain meanings/functions (Aikhenvald, 2014; König & Siemund, 2007; Rett, 2011; 2021). This observation raises the question as to how a particular clause type performs a non-prototypical illocutionary function. In other words, what factors override the prototypical meaning and profile the non-prototypical one? As noted by Comrie (1985), ascertaining how linguistic forms come to develop secondary uses would help to understand the language users’ semantic competence. Following this line of thinking, the current study investigates the phenomenon of mirativity in Urdu and explores how the discourse particle to, a multifunctional lexical item, behaves as a mirative marker in declarative clauses. Addressing this question is likely to help understand the relevant principles of semantic composition responsible for the polysemy often displayed by clause types.

As a category of meaning, mirativity indicates the surprise associated with a speaker’s unprepared mind (Aikhenvald, 2004; DeLancey, 1997). Various studies found that some languages use grammatical categories to encode mirativity, while others use strategies to encode it (Aikhenvald, 2012; Bashir, 2006, 2010; DeLancey, 2001; Peterson, 2017, 2018, 2020). Preliminary research shows that Urdu does not treat mirativity as a grammatical category; rather, it makes use of strategies to encode it. Hence, mirativity in Urdu is parasitic on certain kinds of structure that are not specialized to communicate mirative meaning (Peterson; 2017). To illustrate, Montaut (2006) finds the aorist form of verb as a mirative strategy in Urdu/Hindi. Hook (1974; as cited in Bashir, 2006) observes that “in cases where the performance of an action is completely unforeseen by the speaker, he may not use the compound verb” (p. 248). Bashir (2010) also notes that the compound verb in Urdu/Hindi is incompatible with the semantics of a mirative phenomenon as it fails to signify unexpected new information. Another mirative strategy in Urdu/Hindi is the use of the preterite forms of the verbs thaharnaa ‘to remain, to be’ and nikalnaa ‘to turn out’ (Sigorskiy, 2013).

Recent studies (Bashir, 2010; Bustamante, 2013; Mocini, 2014; Peterson, 2020) suggest that how some languages deploy other grammatical categories for mirativity is yet to be explored. This is also true for Urdu. Urooj et al. (2021), for instance, have discussed a prosodic strategy for emotion realization, but their primary focus is not on exploring the nature of prosody as a mirative strategy. Hussain and Khan (2023) seems to be the only study in Urdu that explores the linguistic realization of mirativity through phonological structure at clause level. Otherwise, there is no study available that concerns the role of the Urdu discourse marker to in mirative meaning construction at clause level.

The discourse marker to in Urdu, akin to but in English (Kehler, 2002), is used to indicate different types of relations. The examples in (1a-1d) below illustrate the polysemous nature of the Urdu to.

(1)  a. əgər  koi   bat     ho         gi     to¹    m   ʊs    se      kəh   ga

    if      any  thing  happen  will  then  I     him  with  say    will

    ‘If something happens, then I will tell him.’    (to ‘then’)

1. meri tərəf to  dek^ho

    my    direction        look 

   ‘Do look at me.’                                              (to ‘do’ as for emphasis)

1. to yeh   tera  kam    hɛ?

    well  this  your doing  is  

    ‘Well, this is your doing?’                                (to ‘well’)

1. əpn g^hər m   to   cəmce   h     k te   nəh

    our   house  in          spoons  are   forks   not   

   ‘There are  only spoons, not forks in our house.’ (to ‘only’)

(Urdu Lughat)

The Urdu to can be used to express what Kehler (2002) refers to as the ‘violated expectation’ relation, that is, counter-expectations in the given discourse. The following examples (2a-2c) express mirative effects.

(2)  a. roṭi      to   jəl     gəi

    bread        burn  went

   ‘The bread got burnt out!’

1. kaṛi to bʊs   gəi

    curry         stale went

    ‘The curry went stale.’

1. pəʊda to mʊrj^ha    gɪa

    plant          wither   went

   ‘The plant withered away!’

The Urdu particle to in (2) indicates, in Fahnestock’s (1983) terms, discontinuative relation (e.g., adversative), such that it signals the lack of common ground between the interlocuters and generates expectations in readers’ minds, accordingly.

The particle to, when placed in pre-subject position, usually serves as a thematic marker in information structure, as in (3).

(3)  to  sirət ne k^hana  k^ha lia

to  seerat   food    ate  took  

‘Seerat ate the food.’

The polysemy of linguistic items, as illustrated above, poses challenges to both translators and lexicographers. Whereas, the native speakers of any language produce various senses in online discourse effortlessly and unconsciously (Ravin & Leacock, 2000). This is because the multiple uses of any lexical item are entrenched in their mental lexicon. Lexical semanticists are also challenged by the phenomenon of polysemy due to the fact that the senses of words cannot be restricted to any particular number (Kovacs, 2011). Considering the polysemous behavior of discourse particle to explicated above, the current study addresses the following question: How do declaratives become miratives in the presence of the discourse particle to? The following section presents the theoretical framework characterizing the nature of conditions that allow a non-mirative declarative clause to function as a mirative expression.

Discourse markers fulfil a non-propositional function, scoping over both sub-sentential as well as supra-sentential units, and tend to be multifunctional. So, the theoretical challenge they pose is to account for how their various uses are interrelated (Hansen, 1998). The received view of meaning, as argued by Evans (2009), isolates a word’s ‘core’ meaning and attributes meaning variations to its interaction with the context. A word’s core meaning is composed of context-independent meaning components that are interpreted through the principles of language use. Sentence meaning stems from the individual or literal meanings of words and speaker meaning depends upon the context of use. The received view of meaning holds that a word retains an idealized, timeless meaning and language users keep this lexicalized meaning apart from its contextualized meanings. As argued by Searle (1983), however, each word’s semantic contribution is a function of its use or context.

Another approach referred to by Pustejovsky (1995) as ‘sense enumerative lexicons’ posits a granular lexicon in which lexical items are assumed to be associated with a vast number of distinct senses. However, positing such a granular lexicon, as Pustejovsky observes, would require an infinite lexicon and given memory constraints, this position is untenable. As stated by Hansen (1998), to understand the contribution of words to meaning, it is not sufficient to assume that in language use, the variation in word meaning is, in part, a function of words, as they serve as access points to encyclopedic knowledge. This knowledge is, then, narrowed by the context of use. In addition, it needs to account for how utterance meaning arises, for which one must also account for how lexical meanings interact with constructional meanings to create the overall linguistic meaning (Goldberg, 1995; 2006). In other words, a descriptive, coherent, and cognitively plausible account of semantic compositionality is required. Such an account should focus on both linguistic and non-linguistic resources to explain how language users deploy words in socioculturally and physically contextualized communicative events (utterances). Against this background, this study deems Evans’ (2009) Lexical Concept and Cognitive Model (LCCM) theory relevant to address its research problem.

Meaning construction in LCCM is based on the interplay between two knowledge systems: linguistic and conceptual. This theory models linguistic system in terms of ‘lexical concept’ and conceptual system in terms of ‘cognitive model’. Lexical concept is the unit of linguistic content associated with a phonological vehicle, while cognitive model is the unit of conceptual content. The semantic value of an expression resides not in the lexical concept or in the cognitive model individually but in the relationship between the two. Each lexical concept in an utterance must observe its unique lexical profile which consists of its semantic and formal selectional tendencies. Semantic selectional tendencies of a lexical concept refer to the other types of lexical concepts with which it can co-occur, whereas its formal selectional tendencies refer to the other types of phonological vehicles with which its phonological vehicle appears. For producing a contextually situated utterance, an open class lexical concept, unlike a closed class lexical one, prompts a cognitive model profile.

Cognitive model profile consists of the experiential knowledge regarding any lexical concept in the mind of the hearer. When the auditory system of the hearer receives any phonological form, multiple types of knowledge come to the hearer’s mind in service of the phonological vehicle. This knowledge, however, is constrained by linguistic as well as extra-linguistic contextual factors. As a result, multiple types of knowledge are narrowed to the contextually appropriate knowledge activated by the particular lexical concept. This process is referred to as lexical concept selection in LCCM theory which provides ground to semantic composition. Meaning variation emerges because of the fact that one form is associated with more than one lexical concept across different contexts. When one linguistic form is used in a particular context by the speaker, the hearer selects one contextually relevant lexical concept rather than the other. Lexical concept selection takes place due to contextual factors which help the hearer to activate the relevant cognitive model in response to the uttered lexical concept. Further, lexical concept selection can take place in two stages: broad selection and narrow selection. In broad selection, the hearer selects one lexical concept associated with one vehicle, while in narrow selection the hearer chooses one parameter within the lexical concept. The phenomenon of multiple selection occurs when a vehicle is associated with more than one lexical concept in the same context.

When relevant lexical concepts are selected with the help of contextual factors, both linguistic and extra-linguistic, the selected lexical concepts are then subject to lexical concept integration influenced by the linguistic context. This cognitive process involves integrating the linguistic content encoded by all lexical concepts in an utterance. Lexical concept integration can be of two types: internal integration and external integration. In internal integration, concepts are integrated with the vehicles and when every vehicle is specified, it undergoes external integration where each vehicle is integrated with other vehicles. Both types of integration are managed by three basic principles. Internal integration is dealt in accordance with the Principle of Linguistic Coherence which states that ‘a lexical concept which is internally open may only be integrated in terms of linguistic content until and unless there is a lexical concept with which it shares schematic coherence’. This principle, in turn, depends on the Principle of Schematic Coherence which states that ‘there should be coherent fusion among the content associated with participants and entities. This principle is further warranted by the Principle of Ordered Integration stating that ‘the process of lexical concept integration in internally open lexical concepts is performed in application to internally simpler lexical concepts before the application to more complex lexical concepts’. Both types of lexical concept integration are influenced by the lexical profile of each lexical concept and result in a lexical conceptual unit. A lexical conceptual unit is the larger unit consisting of different lexical concepts and is interpreted at the final stage of semantic composition called interpretation.

At the interpretation stage, conceptual content is activated, which must cohere with the linguistic content. This semantic compositional process involves both linguistic and extra-linguistic context. The relevant cognitive models are activated in all types of contexts. The next phase is to match these activated cognitive models with each other. This matching, when established, characterizes the utterance in terms of information. The matching of cognitive models with the output of lexical concept integration functions in a way that it is quite compatible with lexical concept integration (the Principle of Guided Matching). At the interpretation stage, matching is constrained by the Principle of Conceptual Coherence which holds that the person, places, and entities encoded by language must exhibit coherence to access relative cognitive models for a unified interpretation. This principle determines how informational characterizations belonging to distinct cognitive models are selected. Only those cognitive models are selected for interpretation which are schematically coherent. Another principle involved is the Principle of Schematic Salience which holds that some conceptions are automatic and the users do not have to make mental efforts for these concepts. When the concepts are overridden by extralinguistic context then different cognitive models are activated simultaneously, which is referred to as the Principle of Simultaneous Activation. Hence, these two principles, namely the Principle of Schematic Salience and the Principle of Simultaneous Activation distinguish between the default interpretation and the distinct interpretation of an utterance.

Research Methodology

This study explores how the Urdu discourse particle to contributes mirative value to a declarative clause type in Urdu. To tap into language users’ intuition about mirativity, the current study requires an inductive analysis of descriptive data (Schütze, 1996). Qualitative research design meets this requirement, as this design can flexibly respond to new details emerging during the investigation (Bogdan & Biklen, 1997; Dornyei, 2007; Gonzalez-Marquez et al., 2007; Robert, 1994). As mentioned above, Urdu employs various linguistic strategies to encode mirative meaning. This study, however, is limited only to declarative clause types that carry expressive content (Zheltova, 2018). The multiple sources used for data collection include a list of mirative expressions elicited from the native speakers of Urdu, TV dramas, movies, and a novel. Urdu Lughat is also consulted and introspection as a source of data is also employed to produce possible utterances. The assumption underlying the use of multiple data sources in this study is based on the rationale that “multi-source evidence can either validate the theory or bring contradictory results, therefore opening new perspectives” (Grisot & Moeschler, 2014, p. 10).

The current study focuses on declaratives without a rising intonational contour and where only the presence of the discourse particle to seems to be the locale of mirativity. This focus leads to the choice of clauses such as X yəh  hɛ ‘X is here’ and X to yəh  hɛ. This type of research focus necessitates the use of the minimal pair methodology to explore how the clauses constituting the minimal pair undergo cognitive semantic compositional processes to suspend their canonical communicative functions and perform mirative functions. This study considers the clausal minimal pair where the contrastive structural distance between the two members is minimal (Gunlogson, 2004). It makes use of such minimal pairs to narrow down the analytical focus on the semantic composition of miratives originating from the minimal contrastive differences and explicitly characterizes the factors responsible for such an interpretation.

Findings and Discussion

In terms of the LCCM theory, the discourse marker to is a symbolic unit characterized by multiple lexical concepts which are the function of the discourse context, as detailed in Section 1 above. This section explicates the mirative contribution of to to the information characterization of the declarative clause əli yəhã hɛ. The following sections explicate how the discourse particle to interacts with other lexical concepts in the given utterance. Further, how it undergoes semantic compositional processes to contribute mirative meaning to otherwise a declarative clause (with the prototypical function of a statement) is also explored. Before accounting for its mirative meaning in the presence of the discourse marker to, however, the conceptual basis of the declarative clause əli yəhã hɛ is in order.

(4)  əli   yəhã   hε

      ali   here    is

      ‘Ali is here.’

The declarative clause əli yəhã hε in (4) above is composed of three lexical items – ali, yəhã, and hɛ – and has the copulative pattern ‘subject (NP) + subject complement (ADVL) + Predicator (VP)’. In terms of LCCM, both the lexical items of the utterance and the copulative pattern in which these items appear assume the status of constructions. əli as noun represents [THING], an autonomous conceptual unit (Langacker, 2008). The Urdu yəhã meaning ‘here’, like its English counterpart here (Huddleston & Pullam, 2005), is a pro-adverb deictically used to direct the attention (Fillmore, 1975; Lyons, 1995) and invoke the idealized cognitive model associated with setting (temporal/spatial references) in the hearer’s conceptual structure (Lakoff, 1987). Its semantic specification is constrained by the Principle of Proximity which allows interpretation with respect to who, where, and when of an event (Williams, 2019). hɛ as a verb symbolizes the [RELATION] conceptual unit. In the context of the utterance above, it relates the subject and its space (physical) and thus represents the relation of location. Being a construction (a form-meaning pairing), the copulative clause pattern itself represents a declarative lexical concept that prototypically encodes statement.

In contrast to the declarative lexical concept that encodes statement, mirative lexical concept indicates a progression from the epistemic state of unknowingness to knowingness, and thus updates the hearer`s cognitive model. Urdu speakers use strategies such as rising intonation and lexical items in context. This contextualised use of strategies enables the hearers to distinguish a declarative lexical concept, the primary concept, from a mirative lexical concept, which is the secondary concept. In the current analysis, the claim is that just like prosodic construction (Sadat-Tehrani, 2008), to – under certain contextual constraints – marks declarative clause as mirative and is in complementary distribution with the prosodic construction. As a symbolic unit, to may be associated with various lexical concepts of which the mirative lexical concept may stand out in certain contexts. As pointed out by Dash et al. (2022), mirativity concept is dependent on the mirative locale and this mirative locale can comprise nominal, verbal, and clause level vehicles. This study aims to explore the linguistic as well as non-linguistic factors involved in suspending the basic elocution of the declarative clause and selecting mirativity as a lexical concept in the presence of to.

As stated in Section 2 above, the lexical concept selection stage of a semantic composition involves both linguistic and non-linguistic contexts. So, it is pertinent to sketch out the context of use first.

Context: Ali’s mother goes to his room to see him but he is not in the room. It is 9 a.m. She thinks Ali must have left for the university. Then she goes to the backyard where she finds Ali sitting with his father. Upon seeing Ali with his father, the mother utters:

(5)  əli   to  yəhã   hε!

       Ali       here   be

      ‘Ali is here!

Ali’s father would take the expression əli to yəhã hε to mean that Ali’s presence over there counters his mother’s expectations. The father makes this inference due to the interaction between the linguistic content and the conceptual content involved in the utterance. The LCCM-theoretic account of semantic composition and the impact of various types of contexts is as follows.

The utterance level context includes the formal features of the utterance, such as its syntactic order and prosodic structure. In the absence of prosodic features such as rising intonation, the word order of the declarative vehicle and the distribution of the particle to assume communicative relevance. The hearer is facilitated by the pre-adverbial position of to which directs his attention towards yəhã as a locale of mirative meaning. In contrast, the non-prosodic utterance əli yəhã hε without to would be taken as a mere statement without any mirative effect.

The discourse level context also plays its role in the selection of the mirative lexical concept. This type of context includes the knowledge discourse participants bring to the discourse event, both textual knowledge and situational knowledge about the usage event, that is, utterance. The discourse level context can be equated with the construct of ‘common ground’, as pointed out by Stalnaker (1973). In the communicative event in (5) above, the common ground may include Ali’s identity and his daily routine, but not Ali’s presence in the given situation. It also includes the discourse participants’ knowledge about the discursive function of to with the existing syntactic distribution. The selection of the mirative lexical concept associated with to is guided by the shared knowledge about to.  Otherwise, the situation might signify a communication failure. Given the same context as in (5) above, if the father says in response:

(6) əli    tʊmhare   samne   bεt^ha  hε   tʊm   ye     kəy˜u    kεh   rəhi      ho   

     Ali    your         front      sit      be you   this   why      say    remain be  

     ‘Ali is sitting in front you. Why are you saying so?’

then the father fails to understand the speaker’s communicative intention behind using to. As observed by Brinton (2010), the discourse remains grammatically acceptable but would be judged “unnatural”, “awkward”, “disjointed”, and “impolite” if such markers are omitted.

In the interaction above to be successful, the participants of the event, the setting of the event, and the encyclopaedic knowledge of the participants also play a significant role as they help the father to associate the mirative lexical concept with to. Given the fact that Ali is sitting with his father, there is no point in saying əli to yəhã hɛ ‘Ali is here!’. The incompatibility between the mother’s present expectation structure and Ali’s actual status makes her use the particle to for the surprising attitude which the father rightly recognizes. After its selection, the mirative lexical concept undergoes integration which is a matter of linguistic context. 

As mentioned earlier, the declarative pattern itself is a construction, that is, a form-meaning pairing.  Its vehicle ‘DEFINITE-NP AdvP be-FINITE’ underlying əli yəhã he encodes [ATTRIBUTION OF A LOCATION TO AN ENTITY] a lexical concept. This lexical concept is propositional as it encodes propositional content. In contrast, the particle to is associated with the [MIRATIVE] lexical concept which is non-propositional in nature, as it encodes no propositional content in it. The integration between the [MIRATIVE] associated with to and the [LOCATION] associated with yəhã takes place under the principles of linguistic coherence and schematic coherence for semantic content and schematic content, respectively. This integration results in the vehicle yəhã indicating a lack of common ground. Linguistic coherence and schematic coherence should work in tandem. This is the reason that the expression *m  janta h  əli yəhã hε ‘I know Ali is here!’ is not acceptable with to since the vehicle ‘janna’ ‘know’ encoding [AWARENESS] does not share the schematic content of the [MIRATIVE]. Information cannot be surprising if it is not new, unknown to the hearer. As required by the formal selectional tendencies, the [MIRATIVE] lexical concept must have a vehicle to specify the mirative locale. In the example utterance, to makes yəhã a mirative locale.

As a result of internal and external integration, the following lexical concepts in the situated utterance əli to yəhã hε! are identified:

(7)  a. əli associated with [THING]

1. to associated with [MIRATIVE]
2. yəhã associated with [LOCATION]
3. hε! associated with [PRESENCE AT LOCATION]
4. e. DEFINITE-NP AdvP be-FINITE associated with [ATTRIBUTION

   OF A LOCATION TO ANENTITY]

At the interpretation stage, the linguistic content resulting from integration activates the relevant conceptual content (cognitive models). Being an abstract concept, the [MIRATIVE] lexical concept takes its conceptual content from the propositional content (Lakoff, 1987). When the matching occurs between the cognitive model accessed by [MIRATIVE] and the cognitive model accessed by the locale of mirativity, it overrides the default reading. When this type of matching takes place, the locale of mirativity yəhã is updated by the surprising attitude about the particular location. This matching aligns with the Principle of Highlighting as a cognitive model is updated when a lexical concept is highlighted. Mirative cognitive model updates the location model under guided matching, since the mother’s cognitive model of Ali does not present Ali sitting with his father. Apart from the Principle of Highlighting, the Principle of Conceptual Coherence and the Principle of Schematic Coherence also govern matching at the interpretation stage. Under these principles, location should occur with the entity that can occupy the location. However, in the presence of [MIRATIVE], the application of these principles remains slightly different. In this case, Ali and his location as his attribute are found mutually incoherent in the mother’s mind and the particle to signals this conceptual incoherence. This incongruity (DeLancey, 2001) overrides the prototypical function of a declarative clause and helps to construct mirative information characterization.  

Conclusion

This study explored the licensing conditions for the Urdu discourse marker to’s mirative contribution to a declarative clause. The data was collected from multiple sources and a representative declarative clause əli to yəhã hɛ! ‘Ali is here!’was selected for analysis. The analysis was framed in terms of the semantic compositional processes as proposed in Evans’ (2009) Lexical Concept and Cognitive Model. The default communicative function of the declarative clause əli yəhã hɛ is a statement, but the presence of discourse particle to transforms it into a mirative expression. The mirative function of to is subject to certain use conditions as it is governed by both linguistic and extra-linguistic context. The study explicates how various licensing conditions such as both linguistic (at utterance as well as discourse level) and extra-linguistic context contribute to the overall information characterization of a clause type. It shows that meaning construction results from the interaction between the two forms of knowledge representation, namely linguistic and conceptual systems.

Conflict of Interest

The authors of the manuscript have no financial or non-financial conflict of interest in the subject matter or materials discussed in this manuscript.

Data Availability Statement

The data associated with this study will be provided by the corresponding author upon request.

Funding Details

No funding has been received for this research.