The Use of Nǐ ‘You’ for Third-Person Reference in Mandarin Conversation

Abstract

In modern spoken Mandarin, the second-person pronoun nǐ ‘you’ is frequently used beyond its canonical addressee-reference function, including for third-person reference. Although previous studies have documented this phenomenon, little is known about how it is sequentially organized, what interactional work it accomplishes, and how recipients orient to it in naturally occurring conversation. Drawing on conversation analysis, this study examines the use of nǐ for third-person reference in the DIG Mandarin Conversations (DMC) Corpus, focusing on its sequential construction, recipient uptake, and interactional import. The analysis shows that this practice typically follows a recurrent sequential pattern: the referent is first introduced in initial position through an unmarked recognitional form, after which nǐ appears in subsequent position, serving as a marked resource for stance-taking. In response turns, minimal tokens function to provide acknowledgment and possible affiliation, whereas sentential responses enable stance modulation and topic redirection. These findings demonstrate that nǐ used for third-person reference is not an incidental deictic shift, but a recurrent interactional practice for the display and management of stance in Mandarin conversation. This study contributes to broader discussions on person reference, footing, and intersubjectivity in talk-in-interaction.

1. Introduction

In talk-in-interaction, person reference is a key resource for organizing participation, social relations, and action (Wortham, 1996; Levinson, 2007). A particularly productive line of research has focused on deictic expressions and their role in indexing participation frameworks (e.g., Sidnell & Enfield, 2016; Stukenbrock, 2020; Holt & O’Driscoll, 2021). Jespersen (1924) termed such context-sensitive linguistic forms shifters, since their referents shift according to changes in the interactional context. Among these forms, personal pronouns are especially consequential for the organization of footing and participant relations (Levinson, 1988; Qu et al., 2021). Prior research on Mandarin conversation has likewise shown that personal pronouns may be used in interactionally flexible and non-canonical ways, with speakers exploiting person shifts to affect recipients’ perspectives and to support the co-construction of discourse (Hsiao, 2011).

In spoken Mandarin, nǐ frequently undergoes deictic shifts whereby it indexes a range of referents beyond its normative use as a second-person singular pronoun (Zhu, 1982). These include first-person referents, second-person plural referents, third-person referents and generic referents (L. Zhang, 2014; Shi & Wang, 2022; J. Zhang, 2022; Luan, 2024). Among these, the use of nǐ for third-person reference is the most common usage in spoken Mandarin (L. Zhang, 2014; Shi & Wang, 2022; Luan, 2024), which highlights its research significance. Data from naturally occurring conversation show ample evidence of this phenomenon. Consider the following excerpt drawn from a televised conversation1 in which Chen and Yu are talking about Yu’s mother. In Chen’s turn, the second occurrence of nǐ is not addressee-directed, but refers to Yu’s mother. The nǐ in Yu’s turn is likewise used for the same third-party referent.

Excerpt (1) (J. Zhang, 2022, Ex. 5a)

Chen:

你

会

以

什么样

的

方式

来

反抗

呢,

nǐ

huì

yǐ

shénmeyàng

de

fāngshì

lái

fǎnkàng

ne,

2SG

will

by

what.kind

DE

way

come

resist

PRT

比如说

故意

的,

我

要

气气

你.

bǐrúshuō

gùyì

de,

wǒ

yào

qìqi

nǐ.

for.example

deliberate

DE

1SG

want

annoy

2SG

‘In what way would you resist? For example, deliberately thinking, “I want to annoy you (referring to Yu’s mother).”’

Yu:

刚开始

有

一段

就是

晚上

故意

不

gāng-kāishǐ

yǒu

yí-duàn

jiùshì

wǎnshang

gùyì

bù

just-begin

have

one-period

COP

night

deliberately

NEG

回家

了,

huí-jiā

le,

return-home

PRT

然后

再

就是

有意识

地,

带

很多

同学

ránhòu

zài

jiùshì

yǒu-yìshi

de,

dài

hěn-duō

tóngxué

then

again

COP

consciously

ADV

bring

many

classmate

到

家里

来,

dào

jiā-lǐ

lái,

to

home-inside

come

这样

你

当

同学

的

面

就

不会

不

zhèyàng

nǐ

dāng

tóngxué

de

miàn

jiù

bú-huì

bù

this.way

2SG

in.front.of

classmate

DE

face

then

NEG-will

NEG

给

我

面子,

gěi

wǒ

miànzi,

give

1SG

face

结果

她

在

同学

面前

还是

会

打

我.

jiéguǒ

tā

zài

tóngxué

miànqián

háishì

huì

dǎ

wǒ.

result

3SG

at

classmate

in.front.of

still

will

hit

1SG

‘At the beginning, there was a period when I deliberately did not go home at night, and then I would consciously bring many classmates home, so that you (referring to Yu’s mother) would not embarrass me in front of my classmates, but it turned out that she would still beat me in front of my classmates.’

A growing body of literature has examined this deictic shift from various perspectives, such as stance-taking and virtual dialogues (Shi & Wang, 2022; J. Zhang, 2022), cognitive pragmatics (Wan, 2019), and discourse and interactional functions (L. Zhang, 2014). However, much of this work relies on written texts or scripted interview data rather than naturally occurring spoken interaction. Moreover, the shifted use is often treated as an isolated utterance-level phenomenon, with insufficient attention to the ways in which it is embedded in unfolding interaction. As a result, we still know relatively little about how the use of nǐ for third-person reference is sequentially organized in real-time conversation, what interactional work it accomplishes, and how recipients orient to it.

To address this gap, this study draws on naturally occurring mobile phone conversation data from the DIG Mandarin Conversations (DMC) Corpus (Yu et al., 2024), and adopts a conversation analysis approach, allowing for fine-grained examination of naturally occurring talk-in-interaction. It seeks to investigate how the use of nǐ for third-person reference functions as an interactional resource in Mandarin conversation, focusing on (i) how this practice is sequentially organized in talk-in-interaction, and (ii) how it shapes stance and alignment relations among the speaker, the referent, and the recipient. As the analysis will show, the use of nǐ for third-person reference recurrently emerges in a particular sequential environment and functions as more than a simple alternative to third-person forms. Rather, it reworks third-person reference into an interactionally involving practice that supports stance display and makes recipient uptake consequential, whether through affiliation, mitigation, or redirection. In this way, the study contributes to research on person reference, footing, and stance in naturally occurring interaction, while also bearing on broader cross-linguistic questions of interactional indexicality.

2. Previous Studies on the Shifted Uses of Nǐ

When Jespersen (1924) first introduced the concept of shifters, he used the term in a broad sense to refer to linguistic forms whose referents shift depending on the interactional context, including the basic and normative usage of deictics. For example, I refers to whichever speaker occupies the speaking role at a given moment. While this is a typical and unmarked usage of the personal pronoun I, it still qualifies as a shifter under Jespersen’s definition, because its referent is context-dependent. This broader understanding has continued to inform subsequent work in Western linguistics, where shifters have been studied in relation to deictic interpretation, imagined communicative scenarios, footing, and person shift across languages (e.g., Talmy, 1995; Wortham, 1996; Whitman, 1999; Ahn & Yap, 2023).

In Mandarin Chinese, however, the notion of shifter has more often been used in a narrower sense, namely to refer to cases in which a pronoun indexes a referent other than its canonical one (Wan, 2019). The present study adopts the term in this narrower sense. More specifically, it focuses on cases in which the second-person singular pronoun nǐ indexes a third-person referent other than the current addressee. Early studies by Zhao (1968), Lü (1980, 1985), and Biq (1991) were among the first to note such non-canonical uses of nǐ. In recent years, growing scholarly interest has led to more systematic investigations into this phenomenon,2 including its classificatory patterns, underlying mechanisms, and discourse functions.

One line of research has focused on how shifted uses of nǐ should be classified. An early and influential account is Biq (1991), who distinguished three discourse-functional uses of nǐ: impersonal, dramatic, and metalinguistic. In her analysis, the first two are rhetorical devices to increase vividness, while the metalinguistic use serves as a vocative to get the hearer’s attention. Later studies have more often classified shifted uses of nǐ according to the type of referent involved. Guo (2008), for example, identified three usages of nǐ and nǐn: congruent reference, phantom reference, and discourse marker, though her treatment of shifted uses remained limited. H. Wang (2008) distinguished anaphoric, arbitrary, generic, and exophoric uses; but her discussion was confined to written language and did not explore specific usages in spoken discourse. Drawing on interview data and interactional theory, L. Zhang (2014) provided one of the earliest more systematic accounts of the shifted nǐ in spoken Mandarin, arguing that it may refer to the speaker, participants in communication, third-person referents, people other than the addressee, or anyone. Shi and Wang (2022) subsequently proposed a two-way distinction based on whether the referent includes the addressee. J. Zhang (2022) further refined the standards of classification, arguing that the referent of nǐ should be identified with reference to the minimum conversational scene in which it occurs. Luan (2024) introduced second-person plural reference as an additional category, which had not been included in J. Zhang’s (2022) study.

Based on these classifications, the shifted uses of nǐ in Mandarin can be summarized into four categories: (1) first-person referents (e.g., wǒ ‘I’, wǒmen/zánmen ‘we’) (see Excerpt 2a) (L. Zhang, 2014; Shi & Wang, 2022; J. Zhang, 2022; Luan, 2024); (2) second-person plural referents (e.g., nǐmen ‘you [plural]’) (see Excerpt 2b) (Luan, 2024); (3) third-person referents (e.g., tā ‘he/she/it’, tāmen ‘they’) (see Excerpt 2c) (L. Zhang, 2014; Shi & Wang, 2022; J. Zhang, 2022; Luan, 2024); (4) generic referents (see Excerpts 2d and 2e) (L. Zhang, 2014; Shi & Wang, 2022; J. Zhang, 2022; Luan, 2024; see also Y. Wang (2021) and Tao (2024) for a detailed discussion of generic nǐ in Mandarin Chinese).3 The following excerpts illustrate these shifted uses:

Excerpt (2)

a.	那些	小孩子	闹得	叫	你	不	能	专心	做	事.
	(Chao, 1968, p. 648)
	nàxiē	xiǎoháizi	nào-de	jiào	nǐ	bù	néng	zhuānxīn	zuò	shì.
	DEM	child	fuss-DE	make	2SG	NEG	can	focus	do	work
	‘Those children made such a fuss that you (referring to ‘I’) couldn’t concentrate on my work.’

b.	你	大学生	要	学	好	科学	知识.	(Dong, 2005 Ex. 6)
	nǐ	dàxuéshēng	yào	xué	hǎo	kēxué	zhīshi.
	2SG	college.student	must	learn	well	science	knowledge
	‘You (referring to ‘you college students’) must learn scientific knowledge well.’

c.	可	班组长	毕竟	还是	工人,	到	外面	谁	也
	kě	bānzǔzhǎng	bìjìng	háishì	gōngrén,	dào	wàimiàn	shéi	yě
	but	team.leader	after.all	still	worker	to	outside	who	also
	不	会	把	你	当	官	看待.	(H. Wang, 2008, Ex. 8)
	bù	huì	bǎ	nǐ	dāng	guān	kàndài.
	NEG	will	BA	2SG	treat	official	regard
	‘But after all, the team leader is still a worker—no one outside would treat you (referring to ‘the team leader’) as an official.’

d.	我	觉得	每	一个	人	都	有	机会，	最
	wǒ	juéde	měi	yígè	rén	dōu	yǒu	jīhuì,	zuì
	1SG	think	every	one CL	person	all	have	opportunity	most
	重要	的	是	你	能	不	能够	承受	那种
	zhòngyào	de	shì	nǐ	néng	bù	nénggòu	chéngshòu	nàzhǒng
	world	DE	COP	2SG	can	NEG	can	withstand	DEM CL
	压力. (J. Zhang, 2022, Ex. 7a)
	yālì.
	Pressure
	‘I think everyone has an opportunity; the most important thing is whether you (referring to “everyone”) can withstand that kind of pressure.’

e.	当	她	感到	气恼	时,	常	低下	头,	从	眼镜
	dāng	tā	gǎndào	qìnǎo	shí,	cháng	dīxià	tóu,	cóng	yǎnjìng
	when	3SG	feel	upset	time	often	lower	head	from	glasses
	顶部		逼视	你.	(H. Wang, 2008, Ex. 18)
	dǐngbù		bǐshì	nǐ.
	top		glare	2SG
	‘When she gets upset, she often lowers her head and glares at you (referring to everyone other than the person under discussion) over the rim of her glasses.’

Although nǐ displays a range of shifted uses in Mandarin, previous studies have consistently identified third-person reference as a particularly frequent and analytically significant pattern in spoken Mandarin (L. Zhang, 2014; Shi & Wang, 2022; Luan, 2024). It is therefore this use that forms the focus of the present study.

Another line of research has sought to explain why shifted uses of nǐ occur and what discourse functions they serve. From a cognitive-linguistic standpoint, scholars have attributed the phenomenon to the speaker’s deictic shift in spatial awareness (Y. Wang, 2008), construal of events and emotions in narration (Liu & Li, 2012), and the pragmatic negotiation between natural and discourse deictic systems (Wan, 2019). From a discourse-functional perspective, the phenomenon has been analyzed in terms of empathy (Dong, 2005), subjective viewpoint (Yao, 2012), and stance. L. Zhang (2014), for example, identifies empathy-recruiting, stance-enhancing, and self-distancing as three major functions of the phenomenon. Shi and Wang (2022), working with naturally occurring spoken materials, further argue that shifted uses of nǐ often construct a form of virtual dialogue through which speakers dramatize the scene and work toward the construction of a shared stance. Relatedly, Tao (2024) approaches generic second-person expressions in Mandarin conversation from an interactional perspective, showing that such forms are closely tied to the local management of social action and participant orientation.

Despite these important insights, previous studies still leave a substantial gap. Most existing accounts treat shifted nǐ primarily as an utterance-level semantic, pragmatic, or discourse-functional phenomenon. Much less attention has been paid to how it is sequentially organized and interactionally recognized in naturally occurring talk. Many studies rely on written texts, literary examples, or scripted interview materials rather than everyday conversation. Even when spoken data are considered, extracts are often fragmentary and are not analyzed with systematic attention to turn design, sequential placement, or recipient uptake. As a result, recurrent claims that shifted nǐ recruits empathy, activates shared knowledge, or promotes stance alignment remain largely inferential unless they are grounded in participants’ observable orientations within unfolding interaction.

This gap is particularly striking in the case of third-person reference, which previous studies identify as one of the most recurrent shifted uses of nǐ in spoken Mandarin, yet which remains underexplored from a conversation-analytic perspective. The few studies that engage more directly with interactional data focus either on first-person shifted uses (e.g., Hsiao, 2011) or on generic second-person expressions (e.g., Tao, 2024), leaving third-person shifted nǐ comparatively undertheorized as an interactional practice.

Against this background, the present study adopts a conversation analytic approach to examine the use of nǐ for third-person reference as an interactional practice in everyday Mandarin talk-in-interaction. The theoretical framework draws on CA’s attention to turn-taking, recurrent linguistic formats, and sequential organization (Sacks et al., 1974; Schegloff, 2007). Rather than treating this phenomenon as an isolated pragmatic strategy, it demonstrates that this usage is designed through a recurrent sequential pattern, functioning as an interactional resource through which speakers actively manage stance to a third-party stance object, and recipients, in turn, orient to it and respond in different ways. In doing so, the study aims to provide a more empirically grounded account of person reference, footing, and intersubjectivity in Mandarin interaction.

3. Data and Method

3.1. Data

This study draws on data from the DIG Mandarin Conversations (DMC) Corpus (Yu et al., 2024), a collection of 308 mundane cell phone calls in Mandarin Chinese from Mainland China, with a total duration of 10 h and 4 min and an average call length of 1.96 min. The transcription system used in the corpus is based on Jefferson’s (2004) conventions, adapted to accommodate the specific linguistic features of Mandarin (see Appendix A for the transcription conventions). The recordings capture a diverse range of everyday interactions among participants in different relationships, including customer–service provider, teacher–student, friends and family members, and doctor–patient, among others. This diversity makes the DMC Corpus well-suited for examining the naturally occurring shifted use of nǐ across different interactional settings.

To identify relevant cases, all instances of nǐ in the corpus were first retrieved using keyword-in-context (KWIC) searches in AntConc (Anthony, 2026). Each retrieved instance was then manually examined in its local sequential context to determine whether the referent was a third party rather than the addressee, a generic referent, or an ambiguous one, and double-checked by a second researcher to enhance analytic reliability.

It is important to situate third-person nǐ within the broader set of generic reference resources in Mandarin. Mandarin does not rely on a single dedicated generic pronoun equivalent to French on or German man; rather, it has multiple ways to encode generalized human reference. Y. Wang (2021), for example, identifies rén ‘person/people’, generic nǐ, and null pronouns as major realizations of inclusive generic reference, while wǒmen ‘we’ and tāmen ‘they’ may function as quasi-inclusive and exclusive generic pronouns, respectively. Constructions such as yǒu rén ‘there are people/someone’ can also encode generalized human reference (Yan, 2011). This background helps clarify why a particular token of nǐ in the corpus may be interpreted as generic or third-person reference depending on sequential and referential cues.

A crucial issue is how third-party nǐ can be distinguished from generic or restricted generic nǐ. Cross-linguistic studies have shown that second-person pronouns are frequently used generically, inviting the recipient to insert themselves into a discursive position made available by the speaker (Kluge, 2016), or to view the described situation from the inside rather than from a detached perspective (Jensen & Gregersen, 2016). Restricted generic nǐ is understood as reference to “anybody in this particular situation” (Kluge, 2016), and usually occurs in statements formulating rules, truisms, or general wisdoms (Auer & Stukenbrock, 2018). In such cases, the predication is not primarily about the properties or conduct of a specific person, but about how people in that type of situation generally act or should act. By contrast, third-party nǐ is understood as a use in which the pronoun is anchored in a specific third party who has already been established in prior talk and remains the relevant stance object in the subsequent sequence. To distinguish third-party uses of nǐ from generic or restricted generic uses, the coding followed three diagnostic criteria.

First, we considered referential anchoring and sequential tracking. A token of nǐ was more likely to be analyzed as third-party reference when its interpretation depended on a specific third party already established through a recognitional form and, in some cases, maintained through unmarked subsequent forms in prior talk, such as third-person pronouns or zero anaphora. Such sequential continuity indicates that participants treat nǐ as part of the ongoing reference to a specific third party rather than as an impersonal generic formulation. By contrast, nǐ was treated as generic or restricted generic when it referred to anyone occupying a relevant role or situation, and when no specific referential continuity could be established. Second, we examined the predicate in which nǐ occurred. This criterion follows Jensen and Gregersen’s (2016) observation that, in generic pronoun use, the point of the predication is to illustrate how the world works in general rather than to describe the properties of specific persons or events. Therefore, if the predicate described the actual circumstances, actions, attributes, or responsibilities of a locally established third party, the token was treated as third-party reference. Third, we considered recipient uptake. The standard methodology provided by conversation analysis is to rely on recipient responses for how the shifted pronoun is understood in interaction (Auer & Stukenbrock, 2018). If the recipient’s next turn continued to orient to the locally established person or event, rather than expanding the turn into a general maxim, this was taken as evidence for a third-party reading.

Only excerpts in which the surrounding sequential context unambiguously established third-person reference and on which both authors agreed were retained.4 Excerpts involving turn abandonment, ellipsis, or other sources of referential ambiguity were excluded.5 This procedure yielded a final dataset of seven excerpts containing ten instances of third person nǐ.

The relative rarity of this practice in the DMC Corpus is itself analytically consequential. Unlike studies based on interview-based data (e.g., L. Zhang, 2014; Luan, 2024), where speakers may have greater opportunity to elaborate viewpoints and deploy person reference more flexibly (Hsiao, 2011), the DMC Corpus consists largely of short, task-oriented telephone calls, such as checking deliveries, making appointments, or requesting information. In this setting, nǐ overwhelmingly occurs in its canonical use, referring to the current addressee. Occasional shifts to third-person reference therefore stand out as marked departures from the unmarked default and are especially worth examining.

The collection is necessarily small, given the rarity of the practice in this corpus. This is not incompatible with a conversation-analytic approach, which aims not at statistical generalization but at identifying recurrent interactional practices through detailed analysis of naturally occurring cases (e.g., Schegloff, 1987; Firth, 1995; Toerien & Jackson, 2019). What is crucial is whether participants demonstrably orient to the practice and whether similar sequential environments recur across cases.

3.2. Method

This study adopts a conversation analytic (CA) approach. CA offers a fine-grained analysis of naturally occurring talk-in-interaction, focusing on how participants organize actions sequentially, design turns in response to prior talk, and build shared understanding moment by moment (Sacks et al., 1974; Schegloff, 2007). This orientation is well suited to the present study, which examines how second person pronoun nǐ is used for third-person reference, how this use is sequentially produced, and how it shapes stance and alignment relations among participants.

The analysis draws specifically on CA’s concepts of turn-taking, recurrent linguistic formats, and sequential organization (Sacks et al., 1974; Schegloff, 2007). It tries to unveil the structural features of this usage in talk-in-interaction, which stand independently of the characteristics of individual speakers or the relationship between them. Knowledge of these structural features (or, in Schegloff’s (2007) terminology, ‘generic orders of organization’) is a key part of our competence as social actors, shaping both our own conduct in talk-in-interaction and our interpretation of the conduct of others (Wilkinson & Kitzinger, 2011). The analytical focus is therefore not statistical generalization but a sequence-based qualitative analysis, with the aim of identifying: (a) how the use of nǐ for third-person referents is sequentially established, (b) what actions speakers accomplish through this usage, and (c) how recipients manage stance projected by the shifted use in subsequent turns.

4. Results

This section presents the analysis in three steps. Section 4.1 examines the recurrent sequential environment in which nǐ is used for third-person reference. Section 4.2 analyzes the interactional work accomplished through this practice, with particular attention to stance. Section 4.3 then considers how recipients orient to it in next turns.

4.1. The Recurrent Sequential Pattern in the Use of Nǐ for Third-Person Reference

Research in conversation analysis has been characterized by an increased interest in “social action formats” (Fox, 2007), i.e., recurrent linguistic formats. Schegloff (1996) distinguishes between initial and subsequent position of reference (i.e., whether the reference is being made for the first time or later in a sequence), and initial and subsequent form (i.e., types of expression typically used for first or later reference—e.g., John vs. he). Recognitionals such as names, kin terms and descriptions, are prototypically used in initial position to establish a referent for the recipient, whereas pronouns are prototypically used in subsequent position once that referent has already been made available in the local sequence (Stivers, 2007; Stivers et al., 2007). In unmarked cases, there is thus congruence between position and form: a recognitional is used first, and a reduced form such as a pronoun is used later.

In the present dataset, such unmarked usage can be observed in both initial and subsequent positions. Speakers typically introduce referents with recognitionals and use unmarked reduced forms in subsequent positions. The recurrent pattern is illustrated in Excerpt (3). Excerpt (3) is taken from a conversation between Hao and Fei, who are colleagues. The individuals mentioned, Liu Qi and Xiao Ming, are also their coworkers. The segment shows how a referent is first established through a recognitional form before the shifted use of nǐ occurs.

Excerpt (3) [OUC-DMC-LZH_Mai zaocan_0000-0115]

…

17	Fei:	我	到	了:	(1.2)	刘琪	家	这	边	了.
		wǒ	dào	le:	(1.2)	Liúqí	jiā	zhè	biān	le.
		1SG	arrive	ASP		Liuqi	home	this	side	ASP
		‘I’ve arrived at Liuqi’s place.’

…

24	Fei:	本来-		本来		我	这::	比较-	我		绕
		běnlái-		běnláI		wǒ	zhè::	bǐjiào-	wǒ		rào
		originally		originally		1SG	this	relatively	1SG		detour
		得		有		点儿	远	呀.
		de		yǒu		diǎnr	yuan	ya.
		DE		have		a.bit	far	PRT
		‘Originally—originally I… relatively—I took a bit of a detour.’
25		我	在		局	里头	去	(取)		上	再
		wǒ	zài		jú	lǐtou	qù	(qǔ)		shàng	zài
		1SG	at		bureau	inside	go	(take)		up	again
		返			回 来.
		fǎn			huí lái.
		return			come
		‘I went into the bureau to pick it up and then came back.’

26	Hao:	.hhh	>º哎呀º<	那	问题	是	那	边儿
		.hhh	>ºāiyaº<	nà	wèntí	shì	nà	biānr
		.hhh	INJ	DEM	problem	COP	DEM	side
		现在::n:		年轻
		xiànzài::n:		niánqīng
		now		young
27		人们	我	就-	不	知道	谁	还	能	用,
		rénmen	wǒ	jiù-	bù	zhīdào	shuí	hái	néng	yòng,
		people	1SG	just-	NEG	know	who	still	can	use
		小明	什么	的	我	就	不		放心	呀.
		Xiǎomíng	shénme	de	wǒ	jiù	bù		fàngxīn	ya.
		Xiaoming	whatever	DE	1SG	just	NEG		rest.assured	PRT
		‘hhh Oh no the problem is, over there now… the young people, I just, I don’t know who can still help. As for Xiao Ming or whoever, I just don’t feel assured.’

…

34		(3.4)
35 →	Fei:	>这 (º些º)<	东西	么	你	说:	不	提前:
		>zhè (ºxiēº)<	dōngxi	me	nǐ	shuō:	bù	tíqián:
		DEM (CL)	thing	PRT	2SG	say	NEG	beforehand
		想	好	你	早上	吃	啥.=
		xiǎng	hǎo	nǐ	zǎoshang	chī	shá.=
		think	good	2SG	morning	eat	what
		‘These (things) you know, you should decide beforehand what to eat in the morning.’

36	Hao:	=唉,	你	(不 xie-)	(.)	说	这些	都	迟	了.
		=āi,	nǐ	(bù xie-)	(.)	shuō	zhèxiē	dōu	chí	le.
		INJ	2SG	NEG xie		say	these	all	late	ASP
		¥你		赶快	去	吧¥.
		¥nǐ		gǎnkuài	qù	ba¥.
		2SG		quickly	go	PRT
		‘(Sigh). It’s too late to say all this now. You should just go quickly.’

In Fei’s turn, line 17, the referent (Liu Qi) is introduced in locally initial position through the name Liúqí, rather than through a pronoun, as in ‘Liuqi’s place’ rather than ‘his place’. The shifted use of nǐ subsequently occurs in line 35, after the referent has already been made available in the local sequence. At this point, continued reference through a third-person form or zero anaphora would otherwise be expected. The use of nǐ is therefore marked not because the referent is unidentified, but because an already established third-person referent is reworked through a second-person format.

Excerpt (4) is drawn from a conversation between a father and his daughter discussing the daughter’s thesis defense. In the segment below, the daughter talks about another student involved in the defense process.

Excerpt (4) [OUC-DMC-YYJ_Biye lunwen dabian_0000-0206]

…

27	D:	然后	给	他们	一人	一份儿.	然后
		ránhòu	gěi	tāmen	yì-rén	yí-fènr.	ránhòu
		then	give	3PL	one-person	one-copy	then
		<我	是	排	在	第二个>.
		<wǒ	shì	pái	zài	dì-èrgè>.
		1SG	COP	arrange	at	ORD-two CL
		‘Then gave each of them one copy, and I was placed second.’

28		(0.5)
29	D:	呃::	其实	一开始	有点	紧张,
		è::	qíshí	yì-kāishǐ	yǒu-diǎnr	jǐnzhāng,
		uh	actually	at-start	a.bit	nervous
		‘Uh, actually at the beginning I was a bit nervous.’

30	D:	因为	一-	他	本来	说	(0.2)	要
		yīnwèi	yī-	tā	běn lái	shuō	(0.2)	yào
		because	one	3SG	originally	say		want
		脱稿,	就是	不	让	你	带
		tuōgǎo,	jiùshì	bù	ràng	nǐ	dài
		speak.without .script	COP	NEG	let	2SG	bring
		稿子	上去.
		gǎozi	shàng-qù.
		script	go-up
		‘Because he originally said the speech should be delivered without a script, meaning you weren’t allowed to bring notes on stage.’

31		(.)
32	F:	嗯.
		ǹg.
		mm
		‘Mm.’

33	D:	结果	he,	结果	那个	he,	我们	那个	第一个(.)
		jiéguǒ	he,	jiéguǒ	nàge	he,	wǒmen	nàge	dì-yī-gè(.)
		result	PRT	result	that	PRT	1PL	that	ORD-one-CL
		第一个	同学	上去	之后-
		dì-yī-gè	tóngxué	shàng-qù	zhīhòu-
		ORD-one-CL	classmate	go-up	after
34		(0.6)
35 →		↑就是(.)	底下	老师	其实	根本	就-	他	就
		↑jiùshì(.)	dǐxià	lǎoshī	qíshí	gēnběn	jiù-	tā	jiù
		just	below	teacher	actually	at.all	just	3SG	just
		(0.5)	不不	听	你	讲.
		(0.5)	bù-bù	tīng	nǐ	jiǎng.
			NEG-NEG	listen	2SG	speak
36 →		你	答辩	的	时间	>完全	就是	<<给 她 看>
		nǐ	dábiàn	de	shíjiān	>wánquán	jiùshì	<<gěi tā kàn>
		2SG	defense	DE	time	completely	COP	give 3SG look
37		(.)
38 →		就是	给	她	留出	时间	来:	让	她
		jiùshì	gěi	tā	liúchū	shíjiān	lái:	ràng	tā
		exactly	give	3SG	leave.out	time	come	let	3SG
		看	你	的	论文儿	的.
		kàn	nǐ	de	lùnwénr	de.
		look	2SG	DE	thesis	DE
		‘So then, well, after our first classmate went up, basically, the teachers down there just—they just didn’t really listen to you. Your defense time was basically just for her to look at. It was just leaving time for her to read your thesis.’

39		(.)
40	F:	嗯.
		ǹg.
		mm
		‘Mm.’

41		(0.3)
42	D:	对	哈哈哈(.)	所以	说	她	上去	就
		duì	hāhāhā(.)	suǒyǐ	shuō	tā	shàng-qù	jiù
		right	haha	so	say	3SG	go-up	then
		开始	开始	念	稿子.
		kāishǐ	kāishǐ	niàn	gǎozi.
		start	start	read	script
43		然后	我	一看	她	念,	然后	老师	也
		ránhòu	wǒ	yí-kàn	tā	niàn,	ránhòu	lǎoshī	yě
		then	1SG	one-look	3SG	read	then	teacher	also
		没	管	(.)	¥我	h¥
		méi	guǎn	(.)	¥wǒ	h¥
		NEG	manage		1SG	h
44		(0.3)
45		¥我	就	不	紧张	了.¥	就	想着	那-
		¥wǒ	jiù	bù	jǐnzhāng	le.¥	jiù	xiǎngzhe	nà-
		1SG	then	NEG	nervous	ASP	then	think	so
46		反正	到	我	我	也	念	呗.	((背景 噪音))
		fǎnzhèng	dào	wǒ	wǒ	yě	niàn	bei.	((background noise))
		anyway	reach	1SG	1SG	also	read	PRT	((background noise))
47		(.)
48		然后	我	就(.)	>就	就	就	就	就<
		ránhòu	wǒ	jiù(.)	>jiù	jiù	jiù	jiù	jiù<
		then	1SG	then	then	then	then	then	then
		感觉	就	好多	了.
		gǎnjué	jiù	hǎoduō	le.
		feel	just	much	ASP
		‘Right haha, so when she went up, she just started reading her script. Then when I saw her reading, and the teachers didn’t intervene, I just stopped being nervous. I just thought, well—Anyway, when it got to me, I’d just read too. Then I just felt much better.’

In Excerpt (4), the referent is first established through a descriptive recognitional diyige tongxue ‘the first classmate’ in line 33. Only after this does nǐ appear in lines 35, 36, and 38. Importantly, the daughter later returns to the third-person pronoun ta ‘she’ in lines 42 and 43. This subsequent return to ta treats the earlier shifted uses of nǐ as continuous with the same already-established third-party referent. The sequence thus shows not only that the referent is introduced before nǐ appears, but also that later reference practices sustain that continuity of referent tracking.

In some cases, the initial reference form is followed by unmarked third-person pronouns and zero anaphora in subsequent position, which maintains ordinary form-position congruence. Excerpt (5) illustrates this pattern. Excerpt (5) is taken from a conversation between a brother and sister discussing their mother’s health; B is the brother and S is the sister.

Excerpt (5) [OUC-DMC-XYZ_Shenti zhuangkuang_0000-0448]

…

06	S:	嗯:	他	看	咱妈↑	二一	年	的-	嗯:::	那个
		èn:	tā	kàn	zán mā↑	èr yī	nián	de-	èn:::	nàgè
		mm	3SG	see	1PL.mom.	21	year	DE	mm	DEM
		检查	结果-	他	说-	这	个:	>狭窄<	的
		jiǎn chá	jié guǒ-	tā	shuō-	zhè	gè:	>xiá zhǎi<	de
		check	result	3SG	say	DEM	CL	narrow	DE
		不	是	很	厉害↑,	说	的
		bú	shì	hěn	lì hài↑	shuō	de
		NEG	COP	very	severe	say	DE
		<不	一定::	要:>	进行	手术↑.
		< bù	yí dìng::	yào:>	jìn xíng	shǒu shù↑
		NEG	necessarily	need	undergo	surgery
		‘Mm he looked at our mom’s 2021 mm that check-up results—he said—the narrowing wasn’t very severe. He said [she] doesn’t necessarily need surgery.’

…

39	S:	行.	>我	我	带	她	拍	就	是	了.<
		xíng.	>wǒ	wǒ	dài	tā	pāi	jiù	shì	le.<
		okay.	1SG	1SG	take	3SG	scan	then	COP	PRT
		拍	完	之后:	然后	我	再	发	给	你
		pāi	wán	zhīhòu:	ránhòu	wǒ	zài	fā	gěi	nǐ
		scan	finish	after	then	1SG	again	send	give	2SG
		给	那个	大夫	看	看.
		gěi	nàge	dàifu	kàn	kan.
		give	DEM	doctor	see	see
		‘Okay. I’ll just take her for a scan. After it’s done: I’ll send it to you, so that doctor can take a look.’

40	B:	º嗯.º
		ºèn.º
		mm
		‘Mm.’

41	S:	↑我	估计:	是	可能	又	二一	年	几
		↑wǒ	gū jì:	shì	kě néng	yòu	èr yī	nián	jǐ
		1SG	guess	COP	maybe	again	21	year	several
		年	了,	肯定	是	严重	了↑,	不然
		nián	le,	kěn dìng	shì	yán zhòng	le ↑	bù rán
		year	PRT	certainly	COP	serious	ASP	otherwise
		那	就	不	会	犯	了	?
		nà	jiù	bú	huì	fàn	le	?
		DEM	then	NEG	will	relapse	ASP	Q
		‘I guess probably been since 2021—several years now—[she] definitely got worse, or [she] wouldn’t have recurred?’

42	B:	嗯.
		èn.
		mm
		‘Mm.’

43	S:	嗯.	我	看	着	那个	眼	也	肿↑,	我	感觉
		èn.	wǒ	kàn	zhe	nàgè	yǎn	yě	zhǒng↑,	wǒ	gǎnjué
		mm	1SG	see	DUR	DEM	eye	also	swollen	1SG	feel
	→	你	是	不	是	劳累	的	事,	也
		nǐ	shì	bú	shì	láo lèi	de	shì,	yě
		2SG	COP	NEG	COP	overwork	DE	matter,	also
		是	天天	精神	紧张	的	事.
		shì	tiān tiān	jīng shén	jǐn zhāng	de	shì.
		COP	every.day	mental	tense	DE	matter
		<吃饭	吧>	倒	是	怪	注意	饮食
		<chī fàn	ba >	dǎo	shì	guài	zhù yì	yǐn shí
		eat	PRT	actually	COP	quite	mind	diet
		方面↑.	就	是	休息	不	好,	我	估计
		fāng miàn ↑.	jiù	shì	xiū xī	bù	hǎo,	wǒ	gū jì
		aspect.	just	COP	rest	NEG	good	1SG	guess
		是	这	一	块-	> 一 天天<	给	自己
		shì	zhè	yī	kuài	> yī tiān tiān <	gěi	zì jǐ
		COP	DEM	one	part	one day.day	for	self
		弄	得	那么	紧张	的,
		nòng	de	nà me	jǐn zhāng	de,
		make	DE	so	tense	DE
		我	的	妈	呀?
		wǒ	de	mā	ya?
		1SG	DE	mom	PRT Q
		‘Mm. I see your eye is also swollen, I feel you might be affected by overwork, and constant mental stress. As for eating, you are quite mindful about diet. It’s just the lack of rest. I guess this is the main part—day by day—[you] made [your]self so tense, oh my god?’

44	B:	º嗯.º
		ºèn.º
		mm
		‘Mm.’

In Excerpt (5), the referent is first introduced using the kin term zán mā ‘our mother’ in line 6, followed by the third-person pronoun tā ‘her’ in line 39, an unmarked example of form-position congruence. In line 41, the clauses kěndìng shì yánzhòng le ‘definitely got worse’ and bùrán nà jiù bú huì fàn le ‘otherwise would not have recurred’ contain no overt subject. The missing subject is recoverable from the prior sequence and is therefore analyzed as zero anaphora referring back to the mother. Like third-person pronouns, zero anaphora functions here as an unmarked subsequent referential form. The shifted use of nǐ appears only later, in line 43. This case is especially revealing because it contains both stages of the unmarked referential trajectory before the marked form occurs: a recognitional form first establishes the referent, and the referent is then maintained through a third-person pronoun and zero anaphora before nǐ is used. This makes clear that the shifted nǐ is not doing the work of referent establishment. Instead, it reworks an already available third-person referent at a point in the sequence where an unmarked subsequent form would otherwise be expected.

Across the collection, then, a recurrent sequential pattern emerges in the use of nǐ for third-person reference:

Recognitionals in initial position to establish the referent (name, kin term, or description)(→ Optional third-person pronoun or zero anaphora in subsequent position)→ Shifted use of nǐ in subsequent position referring to the same third-person referent.

This recurrent sequential pattern helps us understand how word selection is done as part of turn design (Sacks, 1995). It appears interactionally coherent and logical: recognitionals are used in the initial position to establish the referent within the common ground, after which the speaker deploys the shifted use of nǐ. The marked form is therefore not arbitrary but is tightly embedded in the sequential organization and interactional framework of the talk.

4.2. Interactional Actions by the Shifted Use of Nǐ

Section 4.1 showed that the use of nǐ for third-person reference is sequentially organized: it occurs only after the referent has already been established in prior talk. The question then becomes what interactional work this marked practice accomplishes at the moment of its deployment.

In conversation, speakers select reference forms from among alternatives, and these selections are consequential for the actions being performed (Wilkinson & Kitzinger, 2011; Sidnell & Stivers, 2012). When speakers select a marked form instead of a default and unmarked one, they signal that they are doing “more” than “just” referring (Stivers, 2007). Therefore, analyzing the interactional actions accomplished through the shifted use helps reveal how speakers deploy person reference as a resource for action formation.

The present data show that the use of nǐ for third-person reference is closely tied to stance-taking. Stance allows speakers to evaluate objects, position themselves in relation to them, and manage alignment, while drawing on shared sociocultural values (Du Bois, 2007). The marked shifted use of nǐ re-perspectivizes third-person talk by recasting the referent in a second-person format, thereby making the stance object more immediate and interactionally available for recipient orientation.

In many cases, the shifted use indexes a negative evaluative stance toward the referent. For instance, in Excerpt (3), the speaker uses nǐ to refer to a person who failed to state earlier what he wanted to eat, which strengthens the complaining action by treating the referent as someone who could be directly held answerable. A similar pattern can be seen in Excerpt (6), a conversation between a grandmother (Nainai) and her granddaughter (Sunnv), in which the grandmother complains about her son’s eating too few buns during a recent visit.

Excerpt (6) [OUC-DMC-LL_Laomuqin de baoyuan_0000-0022]

01	Nainai:	恁	二叔	此	来	喽	吃	喽	没
		nèn	èr shū	cǐ	lái	lou	chī	lou	méi
		2SG	second.uncle	this	come	PRT	eat	PRT	NEG
		有	仨	馍馍,	我	给	你	学	嘞.
		yǒu	sā	mó mó,	wǒ	gěi	nǐ	xué	lei.
		have	three	bun,	1SG	give	2SG	imitate	PRT
		‘When your second uncle came this time, [he] didn’t even eat three buns, I’ll show you how.’

02		(0.7)
03	Sunnv:	那	吃	啥	啦	?
		nà	chī	shá	la	?
		then	eat	what	PRT	Q
		‘So what did [he] eat?’

04		(1.2)
05	Nainai:	光	喝	点儿,	他	不	咋	吃	馍馍儿.
		guāng	hē	diǎnr,	tā	bù	zǎ	chī	mó mó er.
		only	drink	a.bit	3SG	NEG	much	eat	bun
		‘[He] Only drank a little; he hardly eats buns.’

06		(1.8)
07	Sunnv:	嗷::	(0.2)	来	家	没	事	不	饿	吆.
		ao::	(0.2)	lái	jiā	méi	shì	bù	è	yao.
		oh		come	home	no	matter	NEG	hungry	PRT
		‘Oh when [he] was home, [he] wasn’t hungry.’

08		(1.2)
09	Nainai:	哎:↑呀::↓	你	说-	他	说:	“输	针
		āi:↑ ya::↓	nǐ	shuō-	tā	shuō:	“shū	zhēn
		oh	2SG	say	3SG	say	infusion	needle
		啦,	饿	呗?”	(0.2)
		la,	è	bei?”	(0.2)
		PRT	hungry	PRT
10	→	再	不	饿,	你	zeng	大人	啦,	也
		zài	bù	è,	nǐ	zeng.	dà rén	la.,	yě
		even	NEG	hungry	2SG	such	adult	PRT	also
		得	吃	个	馍馍	不,	晌午	?	(0.6)
		děi	chī	gè	mó mó	bù,	shǎng wǔ	?	(0.6)
		must	eat	CL	bun	Q	noon	Q
11		呃	吃	一	块，	人家-	人家	(0.4)	人家	邦华	都
		e	chī	yī	kuài,	rénjiā-	rénjiā	(0.4)	rénjiā	Bānghuá	dōu
		uh	eat	one	piece	others	others		others	Banghua	all
		吃	俩,
		chī	liǎ,
		eat	two
12		我	说	“看	人家	吃	多	些,	看	你
		wǒ	shuō	“kàn	rénjiā	chī	duō	xiē,	kàn	nǐ
		1SG	say	look	others	eat	more	a.bit	look	2SG
		吃	多	些.”
		chī	duō	xiē.”
		eat	more	a.bit
		‘Oh you know, he said, “After the IV, hungry, right?” Even if you’re not hungry, you’re an adult—you should still eat a bun at noon, right? Uh, eat a piece—look at others, Banghua eats two. I said, “Look how much others eat; look how much you eat.”’

13		(1.4)
14	Sunnv:	.hhh	哎呀,	人家	现在	跟	恁	那	不
		.hhh	āi yā,	rén jiā	xiàn zài	gēn	nèn	nà	bù
		INJ	oh	others	now	with	2SG	there	NEG
		一样	啦,	不	是	吃	zeng	些	馍馍,
		yí yàng	la,	bú	shì	chī	zeng	xiē	mó mó,
		same	PRT	NEG	be	eat	so.many	a.bit	bun
15		人家	都	是	少	吃	馍馍	多	吃	菜.
		rén jiā	dōu	shì	shǎo	chī	mó mó	duō	chī	cài.
		others	all	COP	less	eat	bun	more	eat	veg
		不	跟	恁	[一样	啦.
		bù	gēn	nèn	[yī yàng	la.
		NEG	with	2SG	same	PRT
		‘hhh Oh, he nowadays isn’t like you back then, people don’t eat that many buns. They now eat fewer buns and more vegetables, not like you.’

16	Nainai:	[就	是	哎,	你	要	炒
		[jiù	shì	āi,	nǐ	yào	chǎo
		just		PRT	2SG	need	stir
		点儿	(.)	有	肉	的,
		diǎn ér	(.)	yǒu	ròu	de,
		a.bit		have	meat	DE

17		有	那个:	素	的,	他	光	dao-	dao	那个
		yǒu	nà gè:	sù	de,	tā	guāng	dao-	dao	nà gè
		have	DEM	veg	DE	3SG	only	stir	stir	DEM
		菠菜,	他	不	咋	吃	肉儿.
		bō cài,	tā	bù	zǎ	chī	ròu er.
		spinach	3SG	NEG	much	eat	meat-ER
		‘Right, if you stir-fry something with meat, and the vegetarian one: he only eats the spinach and doesn’t really eat meat.’

In Excerpt (6), the referent is first introduced through the kin term nèn èrshū ‘your second uncle’ in line 1, and is subsequently maintained through zero anaphora and the third-person pronoun tā ‘he’. Although nǐ also occurs in nǐ shuō in line 9, this expression functions as a projecting marker that introduces the grandmother’s subsequent complaint, calling the hearer’s attention to the propositional content and marking the speaker’s emphasis (Biq, 1991). The shifted use of nǐ appears in line 10, anchored in the already established referential trajectory. Although the grandmother’s complaint invokes a common norm that an adult should eat properly, this norm is mobilized to evaluate the second uncle’s specific conduct during his recent visit. This is particularly clear in line 12, where the grandmother reports her own prior utterance and nǐ directly targets the second uncle as the person being evaluated, thereby further supporting the third-person reading of the shifted use in line 10. By recasting the referent in a second-person format, the grandmother’s complaint is rendered more immediate and pointed. The son is treated not simply as a described third party, but as a stance object who can be directly evaluated and held answerable within the interactional space. In this way, the use of nǐ intensifies the evaluative stance and makes the speaker’s concern and dissatisfaction more interactionally salient.

However, the interactional significance of nǐ cannot be reduced to the referent alone. This is particularly evident in Excerpt (4). While nǐ grammatically tracks the first student, the stance object is the institutional situation created by the teacher’s inconsistency and inattention. Although reading from a script works in the speaker’s favor, it contradicts the teacher’s prior instruction that the speech should be delivered without notes and suggests that the teacher is not paying attention to the presentation. In this sense, nǐ functions as a resource for re-perspectivizing the described situation, making it more experientially available and recognizable to the recipient (Biq, 1991; Tao, 2024). Had ta ‘she’ been used, the turn would have remained externally descriptive; by using nǐ, the speaker instead directly pulls her recipient into the narrative and refashions these experiences to make them real and recognizable in recipient’s eyes (Hsiao, 2011), formatting the situation as one that can be experientially inhabited and entered. Furthermore, the multiple applications of nǐ across lines 35–38 increase the rhetorical effect of telling (Tao, 2024), bringing the teacher’s inconsistency and inattention more clearly into focus.

In other cases, the use of nǐ is tied more closely to deontic positioning, that is, the display of rights and authority concerning what should be done (Stevanovic & Peräkylä, 2012). This can be seen in Excerpt (7), taken from a routine interaction between a NICU doctor (D) and the father (P) of a hospitalized newborn, concerning the infant’s medical condition. Linguistic practices in institutional contexts are shaped by institutional expectations and can redistribute responsibility (Heritage & Clayman, 2011; Pilnick & Zayts, 2016).

Excerpt (7) [OUC-DMC-ZH_Baixibao hengao_0000-0154]

…

17	D:	[ang. 呃::	但是	孩子	今天	就是	查	这个	血:,
		[ang. e::	dànshì	háizi	jīntiān	jiùshì	chá	zhè-ge	xuè:,
		PRT PRT	but	child	today	just	check	DEM-CL	blood
18		嗯,	现在	这个	白	细胞	很	高,	hang.	已经
		én,	xiànzài	zhège	bái	xìbāo	hěn	gāo,	hang.	yǐjīng
		mm	now	DEM-CL	white	cell	very	high	PRT	already
		报	危机	值	了.
		bào	wēi jī	zhí	le.
		report	critical	value	ASP
		‘Uh, but today the child just had this blood test. Mm, now the white blood cell is very high. It has already hit a critical value.’

…

23 →	D:	嗯,	你	这个	还是	PC	有	这个	宫内
		én,	nǐ	zhè-ge	háishì	PC	yǒu	zhè-ge	gōngnèi
		mm	2SG	DEM-CL	still	PC	have	DEM-CL	uterine
		感染	的	情况.
		gǎnrǎn	de	qíngkuàng.
		infection	DE	situation
		‘Mm, you still seem to be a case of intrauterine infection.’

24		(0.2)
25	D:	嗯↓.
		én↓.
		mm↓
		‘Mm.’

26		(0.5)
27	D:	嗯::	所以	要	给	你	说	一下.
		én::	suǒyǐ	yào	gěi	nǐ	shuō	yíxià.
		mm	so	need	give	2SG	say	a.bit
		‘Mm. So I need to explain it to you.’

28	P:	嗯,	ang:.
		én,	áng:.
		mm	PRT
		‘Mm, ang.’

…

91	→D:	ang,	如果	[说	你	用	用	药	能	下来,
		ang,	rúguǒ	[shuō	nǐ	yòng	yòng	yào	néng	xiàlái,
		PRT	if	say	2SG	use	use	med	can go.	down
		(.)	应该	就	问题	不	大:,	hang.
		(.)	yīnggāi	jiù	wèntí	bù	dà:,	hang.
			should	then	problem	NEG	big	PRT
		‘Ang, if the medicine works for you, there shouldn’t be a big problem.’

92	P:	[eng.
		[eng.
		eng
		‘Mm.’

93	P:	好.
		hǎo.
		good
		‘Okay.’

Here, the referent is first introduced in the locally initial position through the kin term háizi ‘child’ in line 17, and subsequently tracked through shifted nǐ in lines 23 and 91. The infant’s dependent status is treated as closely affiliated with the parent, reflecting the doctor’s deontic stance by allocating responsibility for the child to the parent. Although the conditional clause in line 91 may superficially invite a generic interpretation (Zobel, 2016), its reading is constrained by the local referential trajectory (Schegloff, 2007; Sidnell & Stivers, 2012). Because the infant has already been established as the focal referent, the conditional clause rúguǒ shuō nǐ yòng yào néng xiàlái ‘if the medicine works for you’ projects a possible treatment outcome for this particular infant rather than a statement about patients in general (Krifka, 1995; Matthewson, 2004).

Deontic positioning may also be accompanied by heightened affective involvement. In Excerpt (5), line 43, when discussing their mother’s health, the sister lists cautions and risk factors (e.g., swollen eyes, overwork, mental stress, lack of rest). The use of nǐ here can be understood in relation to Bühler’s notion of “deixis am phantasma”, in which deictic expressions point to absent entities made accessible through memory or imagination (Bühler, 1934/1982; West, 2013). Here, the sister momentarily brings the absent mother into an imagined interactional space and addresses her as if she were present. This projected address treats the mother as the immediate addressee, thereby strengthening the directive force of the advice and displaying a deontic stance that claims entitlement to recommend and evaluate health-related conduct. At the same time, the shifted use increases interpersonal immediacy. Compared with an unmarked third-person pronoun, nǐ brings the mother into a closer interactional perspective, making the advice hearable as caring involvement rather than detached assessment. In this way, the deontic stance is packaged with affective concern, displaying the speaker’s close and urgent care for her mother.

In sum, these cases show that the use of nǐ for third-person referents is closely tied to stance construction, because it re-perspectivizes third-person talk in ways that make a referent or situation more immediate, more accountable, and more available for shared orientation. This practice is not incidental, but rather context-driven and deployed as a resource for addressing contingencies that speakers face at the moment in ongoing interaction (Tao, 2024). Once referential identification has been secured, the shift to nǐ enables speakers to transform third-person talk into a format through which evaluation, concern, or obligation can be more strongly projected. Through this selection of the marked personal pronoun, speakers involve their hearers to orient them to the co-construction of discourse (Hsiao, 2011). This, in turn, makes the recipient’s next-turn conduct analytically crucial, since recipients may ratify, align with, or otherwise orient to the stance configuration projected by the use of nǐ. The next section therefore examines how recipients respond to this practice in unfolding interaction.

4.3. Response Turns

4.3.1. Minimal Uptake and Affiliation

In six of the ten shifted nǐ tokens identified across the seven excerpts, recipients respond to the shifted use of nǐ with response markers that display minimal uptake and affiliation. These include monosyllabic tokens such as én and áng, disyllabic tokens such as duìya ‘right’, as well as combinations of these forms.

When a single monosyllabic token is employed, it typically performs a basic acknowledgment without displaying any marked affect or stance. In Excerpt (5), for instance, the brother responds with a low-volume èn following his sister’s shifted use of nǐ (line 44). This minimal response performs a basic acknowledgment, as evidenced by the fact that within the same excerpt, the brother also uses èn in response to a pronoun (line 40) and zero anaphora (line 42), both of which are unmarked subsequent referential forms. A similar pattern is observed in Excerpt (4), line 40, where the father responds with ǹg to his daughter’s shifted use, which projects a negative evaluative stance.

Where recipients combine multiple response tokens, however, more nuanced stances may be at play. In Excerpt (7), following the doctor’s turn containing the shifted uses of nǐ (line 23), the father responds with a combined token (én + áng) in line 28, and later produces another combination (eng + hǎo) in lines 92–93. These responses are best understood within the institutional context of medical interaction, where asymmetries in epistemic authority are salient. The use of combined tokens here appears to do more than acknowledge: it orients to the doctor’s superior epistemic status and projects a stance of trust and respect (Heritage, 2012).

Affiliation is more visible when recipients respond with stance-marked tokens. Excerpt (8) is a conversation between two friends who are discussing a COVID-19 patient who secretly returned to Shanxi during quarantine.

Excerpt (8) [OUC-DMC-ZXL_Taiyuan fangte_0000-0507]

…

128	F:	[<山西>,	有	一个	可	丢脸	的,	不	是
		[<shānxī>,	yǒu	yī gè	kě	diūliǎn	de,	bú	shì
		Shanxi>	have	one CL	very	shameful	DE	NEG	COP
		在	人家	河南	还是	在	哪儿,
		zài	rénjiā	hénán	hái shì	zài	nǎér
		at	others	Henan	or	at	where
129		还是	在	武汉	呢,	隔离	的	呢,	结果	偷偷
		hái shì	zài	wǔhàn	ne,	gélí	de	ne,	jiéguǒ	tōu tōu
		still	at	Wuhan	PRT	quarantine	DE	PRT	result	secretly
		[跑	回来	了,	被	骂	上	热搜	啦.
		[pǎo	huí lái	le,	bèi	mà	shàng	rè sōu	la.
		run	return	ASP	PASS	scold	up	hot-search	ASP
		‘In Shanxi, there is one shameful person, [he] was in Henan or somewhere else, [he] was in Wuhan, quarantined, but secretly ran back, got scolded and became a trending topic.’

130	K:	[ang
		ang
		PRT
131	K:	是	吗?
		shì	ma?
		COP	Q
		‘Ang. Is that so?’

132

F:

你

说

山西,

好

事儿

不

出

名,

nǐ

shuō

shānxī,

hǎo

shìér

bú

chū

míng,

2SG

say

Shanxi

good

thing

NEG

come

name

就

这

坏

事儿,

哎呀,

[一

出

名

一

个

准.=

jiù

zhè

huài

shìér,

āiya,

[yī

chū

míng

yī

gè

zhǔn.=

just

DEM

bad

thing

INJ

one

come

name

one

CL

sure

‘You say Shanxi, good deeds don’t get known, only bad ones do, ah, it always makes the news.’

133	K:	[(不	是-)
		[(bù	shì-)
		NEG	COP
134→	K:	=那	人	就	有	有	毛病,	你	就
		=nà	rén	jiù	yǒu	yǒu	máobìng,	nǐ	jiù
		DEM	person	just	have	have	problem	2SG	just
		好好	隔着	呗,	哎,	[不	知道	咋
		hǎo hǎo	gé zhe	bei,	āi,	[bú	zhīdào	zǎ
		good.good	keep distance	PRT	sigh	NEG	know	how
		想	的.
		xiǎng	de.
		think	DE
		‘No. That person definitely has problems, you should just keep distance properly, (sigh), don’t know what [he] was thinking.’

135	F:	[对	呀
		[duì	ya
		right	PRT
		‘Right.’

In line 134, the speaker employs the shifted use of nǐ to construct negative evaluative stance toward the referent, treating the referent’s behavior as improper during quarantine. Although the turn conveys a normative and moral evaluation, nǐ here should not be treated as an impersonal generic form. The description yī gè kě diūliǎn de ‘one shameful person’, together with subsequent zero anaphora in the prior sequence, have already established a specific third party, and the use of nǐ is anchored in this locally established referent. The predicate does not emphasize a general maxim in public-health situations; rather, it focuses on evaluating this particular patient’s reported behavior. Here, the use of nǐ contributes to a rhetorical effect of immediacy and camaraderie by inviting the hearer into the speaker’s world view, implying that the hearer also shares the same perspective (Kitagawa & Lehrer, 1990; Biq, 1991).

In line 135, the recipient responds with an overlapping duìiya ‘right’ immediately following the speaker’s shifted use. Given the speaker’s negative stance toward the referent, both the choice of the response token and its immediate, overlapping placement display strong affiliation with the speaker’s stance. This supports earlier research claims that the shifted use of nǐ can heighten empathic engagement, encouraging recipients to affiliate actively (L. Zhang, 2014; J. Zhang, 2022), while also extending those claims by grounding them in the sequential details of recipient response design. This response further supports the third-person reading of nǐ, since it affiliates with the speaker’s negative evaluation of the patient’s specific conduct rather than expanding the turn into a general piece of public-health advice or a behavioral rule. Compared to monosyllabic response words, such discourse markers may be regarded as preferred responses, in that they not only acknowledge but also further the complaining action initiated by speaker’s prior turn by aligning with it.

4.3.2. Mitigation and Topic Shift

In responses constructed with sentential turn construction units, actions beyond acknowledgment become more complex. A recurrent pattern emerges across these cases: rather than straightforwardly aligning with the speaker’s negative stance toward the third-party object, recipients use their response turns to mitigate the affective intensity of the prior talk and perform a topic shift (Yang, 2010; Yu, 2022).

This is first observed in Excerpt (3), where Fei employs the shifted use of nǐ to deliver a negative evaluation of the third-person referent. The recipient, Hao, responds in line 36 with a turn that begins with the modal particle ai, which displays affective involvement and a degree of alignment with the prior evaluative stance. However, the turn proceeds with a self-interrupted utterance (nǐ bù xie…), which indicates a restraint on the emerging negative evaluation through expressive suppression. Hao’s follow-up utterance (‘It’s too late to say all this now. You should go quickly’.) redirects the sequence from evaluation to future action. In this way, Hao’s response mitigates the prior complaint and offers an exit from extended negative stance-taking.

A similar emotion-regulating function occurs in Excerpt (6). In line 10, where the shifted use occurs, the grandmother is taking a negative stance toward the referent, suggesting that he should eat more. Rather than aligning with this negative evaluation, the granddaughter responds by offering an account for the referent’s behavior, explaining that people nowadays tend to eat more vegetables than buns. This account simultaneously mitigates the grandmother’s negative stance and redirects the interactional focus from the referent’s eating too little to changes in dietary preferences. This topic shift is subsequently taken up in the grandmother’s next turn.

The practice of providing an account to mitigate a negative evaluation and redirect the topic recurs in Excerpt (9), in which two friends are discussing the postgraduate entrance examination.

Excerpt (9) [OUC-DMC-LCT_Tiaoji_0000-0234]

…

42	Jie:	哦:	那	理工	也	是-	理工	也	是	可以,
		o:	nà	lǐgōng	yě	shì-	lǐgōng	yě	shì	kěyǐ,
		oh	then	Sci-Tech	also	COP	Sci-Tech	also	COP	fine
43		啥	都	没有,	然后	就	给	你	个	名单,
		shá	dōu	méiyǒu,	ránhòu	jiù	gěi	nǐ	gè	míngdān,
		what	all	NEG.have	then	just	give	2SG	CL	list
44	→	那	我	怎么	知道	你	有没有	暗箱
		nà	wǒ	zěnme	zhīdào	nǐ	yǒu méi yǒu	ànxiāng
		then	1SG	how	know	2SG	have-NEG.have	hidden
		操作	了.
		cāozuò	le.
		manipulation	ASP
		‘Oh then for Sci-Tech University—it’s also ridiculous, there’s nothing at all, and then just give you a list, so how would I know whether you engaged in any hidden manipulation?’

45	Mi:	那-	那	没有	办法,	那	这	就是	人家
		nà-	nà	méiyǒu	bànfǎ,	nà	zhè	jiùshì	rénjiā
		then	then	NEG.have	way	then	DEM	COP	others
46		的	事儿,	而且	这次	不是	说	他们	说
		de	shìr,	érqiě	zhècì	búshì	shuō	tāmen	shuō
		DE	matter	and	this.time	NEG.be	say	3PL	say
		理工	去	了	贼	多	人,
		lǐgōng	qù	le	zéi	duō	rén,
		Sci-Tech	go	ASP	very	many	people
47		就是	因为	有	很多	家长	也	亲自	去
		jiùshì	yīnwèi	yǒu	hěnduō	jiāzhǎng	yě	qīnzì	qù
		just	because	have	many	parent	also	personally	go
		就是	请	>老师	吃饭<	什么	的.
		jiùshì	qǐng	>lǎoshī	chīfàn<	shénme	de.
		just	invite	teacher	eat.meal<	what	DE
		‘Well—there’s nothing to be done; it’s just their business, and moreover, this time they say that a lot of people went to Sci-Tech, because many parents went in person and, for example, invited teachers to meals and things like that.’

48		(1.0)
49	Jie:	哦:	那	就-	那	就	算	了,	那	=就
		o:	nà	jiù-	nà	jiù	suàn	le,	nà	jiù
		oh	then	just	then	just	let.go	ASP	then	just
		算	了.	看 开	点儿	哇.	能.	行	咱
		suàn	le.	kànkāi	diǎnr	wa.	néng	xíng	zán
		let.go	ASP	take.easy	a.bit	PRT	can	work	1PL
		就	再	来	一年.
		jiù	zài	lái	yīnián.
		then	again	come	one.year
50		不行	了:	找	个	干的,	>你	不是<	在	外面
		bùxíng	le:	zhǎo	gè	gànde,	>nǐ	búshì<	zài	wàimiàn
		not.work	ASP	find	CL	job	2SG	NEG.be	at	outside
		上班儿	了	是	咋	的.
		shàngbānr	le	shì	zǎ	de.
		work	ASP	COP	what	DE
		‘Oh then let it go, let it go. Take it easy. If we can, we can try again next year. If we can’t, find a job. Weren’t you already working outside, or what?’

In line 44, Jie uses nǐ to refer to lǐgōng ‘Sci-Tech University’ while expressing frustration about alleged hidden manipulation. Mi responds by both acknowledging the situation’s uncontrollability and offering an account, noting that many parents personally invite teachers to meals. This intervention has an immediate regulatory effect: Jie replies nà jiù suàn le ‘then let it go’ and moves on to discuss future options such as retaking the exam or seeking employment, after which the conversation shifts to Mi’s own work situation.

Taken together, the data show that recipients design their responses in different ways to accomplish different interactional purposes. While minimal response tokens typically provide acknowledgment and, in some cases, affiliation with the speaker’s stance, sentential turn construction units allow recipients to manage and modulate that stance while simultaneously redirecting the interactional trajectory away from extended negative evaluation. The findings presented across Section 4.1, Section 4.2 and Section 4.3 are summarized schematically in

Table 1. Schematic overview of the recurrent sequential organization and interactional functions of third-person nǐ.

Sequential/Interactional Step	Description	Interactional Function	Representative Excerpts
Referent establishment	A specific third-person referent is first introduced in initial position through an unmarked recognitional form (e.g., name, kin term, descriptive NP).	Secures recipient recognition and anchors the referent before the marked use of nǐ occurs.	Excerpts 3–9
Optional unmarked continuation	The referent is sometimes subsequently maintained through an unmarked third-person pronoun, zero anaphora, or another third-person expression before nǐ is used.	Maintains referential continuity and shows that the referent is already accessible.	Excerpts 5, 6, 8
Shifted use of nǐ	Nǐ appears in subsequent position to refer back to the already established third-person referent.	Re-perspectivizes third-person talk and functions as a marked resource for stance display, including evaluative, deontic, and affectively involved positioning.	Excerpts 3–9
Recipient uptake: minimal response and affiliation	Recipients respond with minimal or stance-marked tokens (e.g., èn, ǹg, áng, hǎo, and duì ya).	Provides acknowledgment and, in some cases, displays affiliation or trust/respect toward the stance projected by the speaker.	Excerpts 4, 5, 7, 8
Recipient uptake: sentential response	Recipients produce fuller response turns, often by offering accounts, advice, or next-action proposals.	Enables stance modulation, mitigation, emotional de-escalation, and topic shift.	Excerpts 3, 6, 9

.

5. Discussion and Conclusions

This study has shown that the use of nǐ for third-person reference in Mandarin conversation is not a free or incidental deictic shift, but a systematically organized interaction practice. The analysis shows that speakers recurrently establish the referent through recognitional forms in initial position and subsequently deploy nǐ to refer back to that referent, thereby using it as a resource for stance display. Recipient responses further reveal how this practice is taken up and managed in real-time interaction: minimal tokens typically register acknowledgment and possible affiliation, whereas more extended responses recalibrate the projected stance by mitigating evaluation and redirecting the interactional trajectory.

A notable feature across the cases is that recipients display no observable trouble in understanding this non-canonical use of nǐ. This raises a further question: how is such a marked usage understood so smoothly in the unfolding progression of talk?

From the perspective of cognitive pragmatics, Wan (2019) argues that referent identification is achieved by the participant roles of both sides of conversation in their indicative coordinates. However, the current study tries to discuss the question from a conversation analytic perspective, and the empirical findings would offer insights into the ways in which intersubjective understandings are constructed in and through the medium of talk-in-interaction (Wilkinson & Kitzinger, 2011).

First, the recurrent sequential pattern identified in Section 4.1 provides an interactional framework that preconditions recipient understanding. The use of nǐ for third-person reference occurs only after recognitionals have been established in initial position, and in some cases, further maintained through an unmarked third-person pronoun in subsequent position. This stepwise sequence builds common ground and anchors the referent prior to the deployment of the shifted use, enabling recipients to treat the subsequent use of nǐ as traceable and projectable to the previously introduced referent within the local sequential environment.

The importance of this sequential design becomes clearer once we consider what would happen if the order were reversed. In Excerpt (8), in F’s narrative, a description (yī gè kě diūliǎn de, ‘one shameful person’) is used as an initial form to introduce the referent in line 128. In K’s turn on line 134, a minimal description (nà rén, ‘that person’) is used in locally initial position, re-establishing the referent before the subsequent deployment of the shifted use. The recipient displays immediate understanding, and no repair is initiated. By contrast, in the following heuristic contrast, the shifted form appears before the recognitional expression:6

Excerpt (10) Hypothetical reversal of Excerpt (8)

134	K:	=你	就	有	有	毛病,	那	人	就
		=nǐ	jiù	yǒu	yǒu	máobìng,	nà	rén	jiù
		2SG	just	have	have	problem	DEM	person	then
		好好	隔着	呗,	哎,	[不	知道	咋
		hǎo hǎo	gé zhe	bei,	āi,	[bú	zhīdào	zǎ
		good.good	keep distance	PRT	sigh	NEG	know	how
		想	的.
		xiǎng	de.
		think	DE
		‘You definitely have problems, that person should just keep distance properly, (sigh), don’t know what he was thinking.

In this reversed version, the shifted use occurs before the recognitional expression nà rén ‘that person’, which disrupts the expected referential trajectory and sounds awkward. Without a prior referent being clearly established, nǐ in initial position is more likely to be treated as addressing the recipient rather than the referent, thereby leading to possible confusion or the need for repair. It thus illustrates that such recurrent sequential design informs and shapes the understanding achieved by the turn’s recipient (Wilkinson & Kitzinger, 2011).

Second, as J. Zhang (2022) argues, shifted uses of nǐ occur when the speaker adopts a non-neutral stance, functioning consistently as a stance marker. The present analysis supports and refines this observation by showing that speakers have typically already initiated stance-taking before nǐ appears. The subsequent use of nǐ does not introduce stance independently; rather, it intensifies, sharpens, or re-perspectivizes an evaluative, deontic, or affective orientation that is already emerging. For example, in Excerpt (3), line 24, by saying wǒ rào de yǒu diǎnr yuan ya ‘I took a bit of- a detour’, the speaker expresses his reluctance, laying the groundwork for the subsequent negative evaluation. The use of nǐ then strengthens the complaint by framing the referent as directly accountable. In Excerpt (7), the institutional setting of the hospital, together with the parent-child relationship, already implies the doctor’s deontic stance that treats the child’s condition as the recipient’s concern and responsibility. The shifted uses of nǐ then make this deontic stance more explicit. In Excerpt (5), the speaker’s prior worry about her mother’s health foregrounds an affectively charged environment, and the later use of nǐ inherits and heightens that concern. Therefore, across these cases, nǐ is interpretable because it occurs within a stance-projective environment, and its use serves to intensify or re-perspectivize the stance already in play. As a result, the interaction progresses smoothly without any initiation of repair, indicating that intersubjectivity is achieved and actively maintained.

This study therefore contributes to the literature in two main ways. Empirically, it provides a detailed analysis of a salient yet relatively underexamined practice, the use of nǐ for third-person reference in Mandarin conversation using CA approach. Analytically, it shows how a form that appears to function primarily as a referential device can be systematically recruited for stance management through sequential organization. By demonstrating that the use of nǐ for third-person reference is both preconditioned by prior referential work and consequential for subsequent response design, the study offers new insights into the interactional grounding of person reference, footing, stance, and intersubjectivity in Mandarin talk-in-interaction. More broadly, it also bears on cross-linguistic research on person reference and interactional indexicality.

At the same time, this study has several limitations. First, cases of nǐ used for third-person reference are relatively rare in the DMC Corpus, which limits how far the present findings can be generalized across spoken Mandarin interaction. Future efforts could strengthen the empirical base by incorporating other spoken Mandarin corpora, which will serve to test the trustworthiness of the current findings and interpretations (Titscher et al., 2000). Second, the present analysis is limited to naturally occurring telephone interaction, and the practices observed here may differ in other communicative environments. Therefore, future research scope could examine a wider range of interactional contexts to explore how interactional settings shape person reference strategies. In particular, computer-mediated communication may provide different resources for reference management and stance projection, especially where interaction is asynchronous, text-based, or multimodally layered. Investigating such environments would help determine whether the interactional organization described here is specific to spoken telephone conversation or reflects a more general feature of Mandarin person-reference practices.