On the Motivations and Pragmatic Functions of Cataphora in Natural Conversation

This paper examines the motivations and pragmatic functions of cataphora in natural conversations. It is found that cataphora is primarily motivated by the speaker and the hearer’s asymmetry of knowledge status about the referent. Other factors are also involved, such as interference, word search, and the tension between the principles of Quantity and Manner. In terms of pragmatic functions, cataphora is used for adding information, emphasis, evaluation, and backgrounding


Introduction
Cataphora, also termed "backward anaphora", is "the process or result of a linguistic unit referring forward to another unit" (Crystal, 1997, p. 68). In contrast to anaphoric pronouns which are linked with previously mentioned entities in the discourse, cataphoric pronouns are forward-referring devices, i.e., their function is to anticipate some aspect of the discourse which follows their occurrence. As noted by Greenberg, "In anaphora, there is co-referentiality between the anaphoric substitute and the antecedent which precedes and is therefore identified. If we reverse this, the cataphoric substitute cannot be coreferential at the moment it is used because we do not yet know what it refers to. After the cataphor has been mentioned we can then in retrospect say that it is coreferential or that when the substitute is used it is proleptically cataphoric" (Greenberg, 1985, p. 283).
This may explain partially why cataphora, in comparison to anaphora, has not been extensively explored by linguists, and the distinction between anaphora and cataphora is rarely examined in grammar textbooks. For this, Quirk et al. (1985, p. 351) offer the following explanation, "cataphoric reference occurs less frequently [than anaphoric reference] and under limited conditions. Where it does occur, anaphoric reference is also possible, so that we can equate two synonymous sentences such as (1a) and (1b) in which the positions of pronoun and antecedent are reversed." (1) a. cataphoric: Before he joined the Navy, Gerald made peace with his family. b. anaphoric: Before Gerald joined the Navy, he made peace with his family.
It should be noted that this explanation takes sentence as an isolated linguistic unit and takes no account of the fact that in natural discourse a sentence may not stand alone but occur in co-text and thus may realize some discourse functions. Indeed, some prescriptive grammarians have gone so far as to condemn the practice of cataphora for reasons of clarity and, more blandly, 'good style', as declared by H.W. Fowler, 'the pronoun should seldom precede its principal' (cited from Wales, 1996, p. 37). This has led to problems in the study of cataphora. A case in point is the borrowing of terminology. Sometimes anaphora is used as an umbrella term including both anaphora (referring backward) and cataphora (referring forward) (e.g., Crystal, 1997;Matthews, 1997). For instance, in TG and GB theory, anaphora is assumed to be the "norm". Within this model, cataphora is often conceived of as the counterpart of anaphora and is referred to as "backward anaphora" (Mittwoch, 1983;Reinhart, 1984;Carden, 1982), presumably meaning something like 'anaphora in reverse'. Some other terms like 'backward pronomilization' are adopted by Kuno (1972). However, these terms are quite confusing to anyone who is assuming that the preceding expression is forward-dependent upon the following expression. Another oddity is that, for cataphora, there are no such correspondent terms as anaphor or antecedent in an anaphoric relation. To characterize such a cataphoric relation, cataphoric word (Crystal, 1997) and antecedent (Quirk et al., 1985) are used to denote the preceding expression and the following expression respectively.
For ease of discussion, the term cataphora is adopted in the present study unless otherwise noted, for instance, in quotations. And cataphor and postcedent are employed respectively for the two coreferential expressions in a cataphoric relation, by analogy with terms anaphor and antecedent.

Previous Work on Cataphora
It is not an exaggeration to say that linguistic theory has devoted relatively little attention to cataphora in comparison to its more familiar anaphoric counterparts. The difficulty of finding true examples of cataphora and the commonplace nature of anaphora suggest the marked status of the former and the unmarked status of the latter, which is probably related to the fact that scarce research work is done on cataphora. And most work on cataphora is based on written discourse and less on spoken language. In the following, we present a very brief overview of the literature on cataphora that is relevant to our analysis.
Generally speaking, the few existing works on cataphora are mostly conducted in the field of anaphora when exploration is made on the various constraints governing the distribution and semantic interpretation of different types of nominal expressions. The syntactic approach to cataphora is mainly done within TG and GB framework, where cataphora has been discussed primarily as a syntactic phenomenon by linguists (Langacker, 1969;Reinhart, 1983), who claim that pronouns could only refer to referents higher in the phrase structure diagram. As Xu (1995, p. 4) remarks, "First, it [formal approach] is largely limited to the formal aspects of anaphora. And second, it is largely limited to the exploration of such formal properties of noun phrases within a sentence." Mittwoch (1983) is the first to argue that cataphora is a feature of discourse structure requiring to be analyzed pragmatically, "whatever constraint is here involved is not a rule of sentence grammar but a pragmatic rule for the appropriate use of sentences" (ibid, p. 130). Harris and Bates (2002) examine the clausal backgrounding and pronominal reference, claiming that the subordinate clause where the cataphor occurs functions as backgrounding. Ariel (1990) and van Hoek (1997) investigate cataphora from a cognitive perspective. Guided by her Accessibility Theory, Ariel (1990) claims that distance and unity are the determining factors for the use of cataphora. Based on the concepts of reference point and dominion, van Hoek (1997) argues that cataphora typically involves a significant asymmetry in prominence between the nominal conceptions corresponding to the pronoun and the antecedent and therefore can be explained by the same factors which explain the constraints on anaphora.
The study of cataphora in Chinese has, to the best of our knowledge, been the object of only a small body of research. Chinese linguists hold different views with regard to the existence of cataphora. Some deny the presence of cataphora, such as Wang (1994) and Wang (2006). Others are against this argument (Gao, 2003;Liu, 1997). Based upon the written discourse, Xu and He (2007) make a contrastive study of cataphora between English and Chinese and find that cataphora is used to code background information. Gao (2010) investigates the cognitive mechanism of cataphora in English and Chinese narrative texts. She finds that the rarity of cataphora is due to the higher cognitive cost involved in its interpretation and the cognitive cost in the interpretation of different types of cataphora is in inverse proportion to their distributional pattern. The prominence asymmetry between the cataphor and its postcedent instructs the reader to construe the zero pronoun or pronoun in the initial position of the sentence as the current topic or as a signal for the termination of the current topic and the start of a new one.
A brief review of the literature on cataphora reveals that the previous studies on cataphora uncover the structures, functions, and cognitive mechanisms of cataphora in written discourse. Whether the conclusions are applicable to spoken language is yet to be testified. The present study will examine the motivations and pragmatic functions of cataphora in natural conversations.

Motivations of Cataphora in Natural Conversation
The cataphoric instances for the analysis below are selected from three corpora that are available on the Talkbank website (https://ca.talkbank.org/). They are SBCSAE (the Santa Barbara Corpus of Spoken American English), LDC (The Call Home Telephone Speech Corpus for American English, Linguistic Data Consortium), and CABNC (Spoken Language Component of the British National Corpus).

Asymmetry of Knowledge Status
In everyday conversations, interactants are interconnected with each other through common ground, which is believed to prompt the speaker to employ a pronoun to refer to an entity introduced into the conversation. After producing the pronoun, however, the speaker may realize that the common ground might be inactive to the hearer at the moment of speaking or the referent coded by the pronoun might be brand new to the hearer. The pronoun therefore could be misleading. Thus, to remove the potential referential obscurity, the speaker will then use a more informative expression, i.e. postcedent, to substitute the cataphoric pronoun.
(2) 01 A: I wrote up the specification I drew up & uh oh about thirty-four pages 02 of very detailed flow charts. In (2), the cataphor "he" in line 04 indicates that speaker A takes the referent coded by the pronoun "he" as given for the hearer. After producing the pronoun "he", however, A realizes that the referent may not be shared with the hearer, thus making the shift from a pronoun to a definite expression "the guy that I ga-gave" in line 06 to make the referent identifiable.
( 3) In (3), speaker A believes that the referents "They" to be introduced in line 01 are in their common ground, thus producing a pronoun "They". Subsequently, she comes to realize that the hearer may be unclear about what the pronoun "They" refers to. Consequently, speaker A clarifies with a more informative expression "computers" at the end of line 01.

Interference
Interference is another factor motivating cataphora, i.e., there are some other potential candidates in the previous sequences competing for coreference with the cataphoric pronoun.
(4) 01 A: Lida and Irv all felt she was just a great mother… In (4), "Lida and Irv" in line 01 is the interference for the referent coded by "They" in line 02. Realizing that, speaker A reintroduces the referent with a more informative expression "Elsa and her husband" for clarification at the end of line 02. It should be noted that the sentence structure "that's X" used by speaker A also indicates that a potential candidate does exist in the prior sequence.
(5) 01 A: And I have these three third-graders.
In (5), speaker A produces the pronoun "they" initially in line 03. Subsequently, A recognizes that there is a potential candidate "these third-graders" in line 01 for the referent of the pronoun. Therefore, A has to produce a definite expression "the teachers" for clarification in line 04.

Word Search
Utterances are unplanned in online interactions. Thus, word search frequently occurs in everyday conversations.
In the course of word search, the cataphor serves to signal that there are difficulties in establishing the intended referent and also as a cue to the hearer that a new referent is about to appear.
In (6), speaker A utters the pronoun "He" twice in line 01. "He" here acts as a mental placeholder, signaling that the referent coded by the pronoun is to be established. The pause indicated with (..), hesitations "uh uh uh", and the prolonging of the sound marked with the colon in "a:nd" after "he" show that speaker A may come across some difficulties to retrieve the name of the referent and has to search his memory for it. Finally, A succeeds with the proper name "Duchamp".
(7) 01 A: and they and the readings that &uh Jack had picked out 02 were just great.

B: yeah. (LDC)
In (7), speaker A uses the pronoun "they" to set the stage for the referent. Thus, the hearer is given a signal from the placeholder that the subsequent referent is forthcoming. By repeating the conjunction "and", A manages to retrieve the referent "the readings" he intends to introduce into the conversation.

Tension Between the Principles of Quantity and Manner
The Cooperative Principle (CP) proposed by Grice (1975) embraces four sub-principles, namely, Quantity, Quality, Relation, and Manner, as illustrated below (ibid: 45).
(8) Quantity 1: The contributed expression should be as sufficiently informative as demanded for the current interaction; Quantity 2: Superfluous information should not be supplied; Manner 1: avoid obscurity; Manner 2: avoid vagueness; Manner 3: be brief; Manner 4: be organized.
Q1 indicates that the interactants are required to supply sufficient information for the present exchange, whereas Q2 specifies that information should be confined to an efficient expression, which is a contradiction. Tension also exists in the maxims of Manner. M1 and M2 suggest that the information provided by interactants should be adequate and unambiguous. M3 and M4, however, delimit that the contributed information has to be brief and organized. As Geluykens (1994, p. 10) points out, there is indeed a tension between these two principles of Quantity and Manner. Consequently, the balance (i.e., efficient information) should be searched for.
In the course of referring forward, cataphora goes through two stages, i.e., the production of cataphor and that of postcedent. In the first stage, the speaker zeroes in on the economy of the information, thus producing the cataphor. After that, the speaker recognizes that it may be difficult for the recipient to identify the referent coded by the cataphoric pronoun. Therefore, a more informative expression, i.e., the postcedent is used by the speaker for clarification in the second stage. We will take the following instances for illustration. It can be observed from (9) that speaker A's referential choice of "she" in line 01 is made on the ground of Q2, M3, and M4. After producing the pronominal form, A recognizes that he has violated the principle and maxims of Q1, M1, and M2. Subsequently, the postcedent "Kirsty" is delivered in line 01. In (10), speaker B produces the pronoun "It" in line 06 to denote a specific entity. Noticing that there are two potential competitors, i.e., "the clutch" in line 01 and "the accelerator" in line 03, speaker B disambiguates by using a definite expression "the foot control", thus observing Q1, M1, and M2.

Clarification
As exemplified by the cataphoric instances above, the cataphor takes the form of a pronoun with inadequate information, thus posing a potential obstacle for the recipient to catch the referent coded by the pronoun. To make the referent explicit, clarification has to be made subsequently.
(11) 01 A: I dare say somebody's already said it to you but the family were very 02 comforted and grateful for your service. In (11), after employing the pronoun "it" in line 01, speaker A realizes that it might be difficult for the recipient to identify the referent. Therefore, A clarifies the cataphor immediately in lines 01 and 02 through a clause "the family were very comforted and grateful for your service". In so doing, obscurity is avoided and the conversation proceeds with smooth progress.
(12) 01 B: My other two companions speak much better Chinese than I do.
02 One of my companions will be finished next May.
03 B: the other is just a gifted person and, 04 you know just picks up really well so you know like 05 A: How long has she been here the one that's gifted?
In (12), noticing that the intervening referent in the prior sequence "one of my companions" in line 02 might compete for the referent coded by "she" in line 05, A uses the definite expression "the one that's gifted" for clarification.

Emphasis
It is found that in cataphora, the speaker not only lays emphasis on some unusual features of the referent coded by the cataphoric pronoun; but also underscores his or her stance on the referent.
In the previous sequences of (13), Speaker B mentioned that "Jon Ward" was under harsh living conditions. B brought the referent "Jon Ward" into focus through cataphora to show his respect for the referent's attitude towards the unpleasant situation.  (14) is taken from the context that the interactants are holding a welcome ceremony for the newly elected Archbishop of Wales. After speaker A produces the pronoun "his" in line 02, he finds that the pronoun is insufficient to put emphasis on the referent's identity as the Archbishop. Thus, the postcedent in lines 02-04 is used to accentuate his sincere respect for "the most reverent Alwyn who has just recently been elected as the Archbishop of Wales". By employing the definite expression with the relative clause, A is attempting to emphasize the identity of the referent.

Evaluation
It is observed from our data that cataphora can also make evaluations. The speaker can not only show his or her own emotions of the referent indicated by the cataphora, but also expect the recipient to identify and agree with the assessments. Thus, relations of alignment are constructed by shared emotions and feelings (Martin & White, 2005, p. 2).
02 B: Yeah but she 03 A: You can't you can't really get mad with her cos she's so nice 04 She never er never nasty with you, is she?
In this example, cataphora is used to evaluate the girl "Bev", who is referred to initially with the pronoun "She", followed by the postcedent "Bev" to make the positive evaluation accountable.
Cataphora can also be used to present negative assessments as illustrated in (16). Line 01 shows that the cat is touching speaker A. This misbehavior arouses A's complaint in line 03, which is expressed with the cataphoric pronoun "her" and the postcedent "silly little cat".

Backgrounding
Backgrounding is to mark less important information in narration (Li, 2018, p. 1). There exists a close tie between backgrounding and dependent clauses as demonstrated in Indo-European languages (Tomlin, 1985). The participles, adverbial clause, and prepositional phrase are strong background markers at the initiation of a clause, implying background information such as time, place, or manner (Khalil, 2005, p. 8 09 A: Cos when they pulled into a slow section on the railway the vandals 10 have got on and just gone through the whole lot and taken the radios 11 out. In (17), the cataphor "they" in line 09 appears in the adverbial clause introduced by "when", indicating the background information of time. The main clause, however, projects crucial information, updating the recipient's knowledge status by uncovering the process of the current events. It should be noted that backgrounding is not always displayed by syntactic subordination. Example (18) In (18), the interactants are talking about Michael's pneumonia. The main clause containing the cataphor "he" carries the background information. The adverbial clause with the postcedent, however, serves as the foregrounding information. This is what Cristofaro (2005) claims "pragmatic subordination", realized by a cataphor-containing clause.

Conclusion
This paper elaborates on the motivations and pragmatic functions of cataphora. In terms of its motivations, cataphora is prompted by the asymmetry of knowledge status, interference, word search, and the tension between the principles of Quantity and Manner. As for pragmatic functions, it is found that cataphora in natural conversations performs four functions, namely, clarification, emphasis, evaluation, and backgrounding. The present research may give some insights into a comprehensive understanding of cataphora and shed more illuminating light on referential study.