Polyfunctionality of ‘Give’ in Hui Varieties of Chinese: A Typological and Areal Perspective

Abstract

The morpheme ‘give’ is among the most well-studied lexical items in the realm of grammaticalization. This study sets out to provide a typological and areal analysis of the distinct forms and multiple functions of ‘give’ in 27 varieties of Hui Chinese, a lesser-known group of Sinitic languages. Making use of both primary and secondary data, we have identified ten different functions of GIVE, namely (i) lexical verb ‘give’, (ii) recipient marker ‘to’, (iii) benefactive marker ‘for’, (iv) purpose marker, (v) permissive marker, (vi) passive marker, (vii) pretransitive disposal marker, (viii) allative marker, (ix) locative marker ‘at/in’, and (x) temporal marker ‘till’. The Hui varieties covered in this study generally showcase the syncretism of a minimum of five of the functions above simultaneously. Semantic extension, polygrammaticalization, and cooptation are shown to be the major mechanisms behind the polyfunctionality or polysemy sharing of the morpheme ‘give’. Our study contributes to the understanding of the role that grammaticalization, especially contact-induced grammaticalization, plays in forming linguistic areas. In addition, it casts doubt on the basicness of ‘give’ in assessing the genetic relatedness of languages in the world.

1. Introduction

Grammaticalization is conventionally defined as ‘the change whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions, and, once grammaticalized, continue to develop new grammatical functions’ (Hopper and Traugott 2003, p. 18). Although grammaticalization is traditionally studied as a language-internal phenomenon, new findings from a number of fields, particularly areal typology (Bisang 1996; Matisoff 1991; Ansaldo 1999; Dahl 2004; Enfield 2003; Heine and Kuteva 2003, 2005), bilingual development (Matthews and Yip 2009; Szeto et al. 2017, 2019), and creole studies (Arends et al. 1994), have provided abundant evidence to demonstrate that grammaticalization could result from language-external factors such as geographical clustering and substrate influence. In their seminal work on grammaticalization, Hopper and Traugott (2003, p. 230) also acknowledge the significant connection between language contact and grammaticalization, concluding that ‘Contact has been an important factor for most languages, and a strictly monogenetic view of grammaticalization is ultimately inappropriate’.

The morpheme GIVE is among the most well-studied lexical items in the realm of grammaticalization. Cross-linguistically, the lexical verb ‘give’ exhibits an array of distinctive functions, as identified in Kuteva et al. (2019, pp. 192–203) (

Table 1. The multiple functions of ‘give’ in the world’s languages (Kuteva et al. 2019, pp. 192–203).

give >		Functions
	(1)	benefactive
	(2)	causative
	(3)	cause
	(4)	concern
	(5)	imperative
	(6)	passive
	(7)	patient
	(8)	permission
	(9)	purpose
	(10)	recipient

).

More relevantly, Bisang (2015, pp. 137–39) identifies four different functions of ʔaoy ‘give’ in Khmer as a coverb, a causative verb, an adverbial subordinator for purpose or manner, and a complementizer. Yap and Iwasaki (2003) focus on a particular grammaticalization cline of the lexical verb ‘give’, from the permissive–causative function to the passive function in many East and Southeast Asian languages, and identify a case of contact-induced grammaticalization in Kedah Malay. Matthews and Yip (2009) further confirm the high likelihood of this case of contact-induced grammaticalization, with data from Cantonese–English bilingual acquisition of the different functions of the lexical verb ‘give’, and the areality of such a pathway in Southern Sinitic varieties.

Regarding Sinitic languages, Tsao (1988), Lai (2001), Chin (2011), Ngai (2015), and Lu and Hui (2023), among others, have provided comprehensive insights into the polyfuncationality of the morpheme GIVE in varieties south of the Yangtze River, namely Taiwanese Southern Min, Hakka, Cantonese, Shaowu, and Tunxi Hui, respectively. These studies reveal that the morpheme GIVE serves six to nine concomitant functions in these Sinitic varieties, with all of them sharing the function of a passive marker, suggesting a pattern of areal distribution induced by language contact.

In light of the complexity and the possibility of areal distribution of functions of the morpheme GIVE, this paper sets out to provide a micro-typological study on a group of lesser-known Sinitic languages situated in the central transitional region of China (Chappell 2015; Szeto and Yurayong 2021). By investigating the forms and functions of the morpheme GIVE in 27 datapoints of Hui Chinese, we aim to find out the following:

(i)

What are the forms of the lexical verb ‘give’ in each of the datapoints? Why should they differ so greatly from each other?

(ii)

What are the functions of the morpheme GIVE in each datapoint of Hui Chinese? What is the mechanism of such radical polyfunctionality of ‘give’?

By answering these questions, we hope to shed light on the mechanism of grammaticalization of GIVE cross-linguistically, and further enrich our understanding of the lexeme GIVE, one of the most ‘basic’ lexical items in human cognition.

2. Data Collection and Hui Chinese

In this study, 27 samples of Hui Chinese (Figure 1)1 are assembled, including primary and secondary data, covering all the five subgroups of Hui scattering over Southern Anhui Province, Northeastern Jiangxi Province, and Western Zhejiang Province (Figure 2) south of the Yangtze River. In collecting primary data, both linguistic elicitation and natural linguistic recordings are used. Unless otherwise indicated, all examples presented in this study are primary data collected from the field. Data from the literature are glossed and translated by the authors if no such gloss or translation is given in the original work.

Before delving into the study of the morpheme GIVE, we would like to provide a brief introduction to Hui Chinese. As one of the least-studied groups of Sinitic languages, Hui Chinese varieties (circled in red in Figure 2) are mainly spoken in the Lower Yangtze River area of China, including Southern Anhui Province, Northeastern Jiangxi Province, and part of Western Zhejiang Province, with approximately 3.2 million speakers (Zhao 2005). In the literature, Hui ‘dialects’ used to be classified either as a subgroup of Mandarin, e.g., Xiajiang Mandarin (Li [1937] 1973), Huining Mandarin (Wang 1955), or Jianghuai Mandarin (Zhan 1981), or a subgroup of Wu Chinese (Cao 2002; Zhao 2004). It was not until the first edition of the Language Atlas of China (Wurm et al. 1987) that Hui Chinese was categorized as an independent group of Sinitic with five subgroups. In the second edition of the Language Atlas of China (Zhang 2012), Hui Chinese retains its status as an independent group of Sinitic languages, albeit with a refined grouping of subgroups, namely Xiu-yi, Ji-she, Jing-zhan, Xi-wu, and Yanzhou (Figure 3).

The history of Hui Chinese can be traced back to bai-yue 百越 ‘hundreds of Yue tribes’. In Chinese historiography, bai-yue conventionally refers to non-Sinitic tribes who are believed to be ancestral to the present-day speakers of Tai-Kadai, Hmong-Mien, and Austroasiatic languages (Li 1994; Meacham 1996; LaPolla 2001). The ancient residents of the present-day Hui region belonged to a particular branch of bai-yue known as shan-yue 山越 ‘mountain Yue tribes’; they were probably either ancestral inhabitants of the Hui regions or had reached northward from the Far South to the Yangtze River and resided in the chains of undulating hills in the Hui regions in ancient times. Such a hilly geographical condition contributes to the substantial level of internal diversity observed between subgroups, which is further reinforced by continuous waves of war-induced, disaster-driven, or government-initiated migrations since the West Jin Dynasty, especially from the Central Plains proper (Coblin 2002; Meng 2005). It was manifested in Huizhou Fuzhi 徽州府志 ‘A history of the Huizhou Provincial Capital’ in the Ming Dynasty that 六邑之語不能相通 ‘The languages of the six counties under the Huizhou Provincial Capital are mutually unintelligible’.

Despite the observable diversity, there are still features that ‘unite’ Hui varieties as a group and distinguish themselves as transitional varieties, with both intermediate features between Northern and Southern Sinitic languages, as well as unique features found only in Hui Chinese. Zhengzhang (1986, pp. 13–14) and Zhao (2005, pp. 279–82) summarize the phonological characteristics of Hui varieties that categorize themselves as distinct from neighboring groups of Chinese as follows. Firstly, all voiced consonants in Medieval Chinese have become devoiced. A two-way contrast is present between voiced and voiceless aspirated and voiceless unaspirated affricates and stops in most varieties of Hui. In other words, the three-term distinction between voiced aspirated, voiceless aspirated, and voiceless unaspirated affricates and stops, as demonstrated in the adjacent Wu Chinese varieties, is not shown. Secondly, rhymes with the nasal coda /ŋ/ or with the nucleus ending with /i/ or /u/ have gradually dropped their codas. Thirdly, rising tones with voiced initials have tended to gradually change into lower rising tones with voiceless initials. Finally, regarding tones, Hui varieties record an average of six tones, mostly with a glottalized checked tone, albeit somehow weakened. Among these features, the number of tones demonstrates the transitionality of Hui Chinese, as Northern Sinitic languages generally have three to four tones, except for some extreme cases, e.g., Wutun, a radically restructured variety of Northwest Mandarin, which has lost all of its tones due to contact with Amdo Tibetan and Bonan (Sandman 2016); meanwhile, Southern Sinitic languages generally have six to nine tones, including checked tones, as in varieties of Yue and Min. The existence and distribution of checked tones are yet another indication of Hui’s transitionality. The checked tones, literally syllables ending with stops /p/, /t/, and /k/ in Medieval Chinese, are fully preserved in Southern varieties like Yue and Hakka, but totally lost in Northern Mandarin. Hui Chinese, on the other hand, generally manages to preserve a merged version of the final stops, a glottalized /ʔ/, although the degree to which it is preserved varies across different varieties.

Morpho-syntactically, Hui Chinese displays ‘transitional’ features intermediate between Northern and Southern Sinitic languages, as well as features unique to itself. On the one hand, regarding the comparative constructions, Hui varieties generally take after the Northern type of the ‘Compare’ comparative constructions, i.e., [Marker-Standard-Adjective], in contrast with the Southern type of what Ansaldo (1999) terms as the ‘Surpass’ type, or in Chappell’s (2015) term the ‘Action’ type. On the other hand, unlike Northern Mandarin varieties, Hui has retained a remarkable number of monosyllabic words from Medieval Chinese, e.g., ɕi¹¹ ‘play’ and uʔ⁵ ‘house’ in Tunxi Hui, which are either no longer used, or have been disyllabicized in Standard Mandarin. A head-initial word order tendency is widely observed with respect to the relative word order of animal names and their genders [animal + gender], as well as the verb and the modifying adverb [verb + adverb], resembling those of many Southern Sinitic languages, e.g., Cantonese. Nevertheless, this tendency has been gradually replaced in recent years with the opposite head-final tendency, in co-existing patterns [animal + gender] or [gender + animal], or double-marking constructions, e.g., [ʦə⁴² ‘again’ ʨʰiʔ⁵ ‘eat’ i^11-5 ‘one’ u:ə³¹ ‘bowl’ tʰiɛ¹¹ ‘again’], under the overwhelming influence of Standard Mandarin. In addition, as has been pointed out in Lu (2018) and Lu and Hui (2023), some Hui Chinese varieties such as Tunxi Hui employs the ‘give’-passive, generally regarded as a feature of many Southern Sinitic and Mainland Southeast Asian languages (Hashimoto 1988; Yap and Iwasaki 2007; Chappell 2015; Szeto 2019, among others). However, given the complexity and internal diversity of Hui Chinese, it is worthwhile to explore more Hui varieties on the lexical sources of passive markers.

After having provided a sketch of the typology of Hui Chinese, our attention will shift to examining the unique forms and functions of a specific morpheme, namely GIVE, in various varieties of Hui Chinese.

3. The Distinct Forms of GIVE in Hui Chinese

Although belonging to the same ‘dialect group’ of Sinitic languages, the ‘basic’ morpheme GIVE of different varieties of Hui Chinese varies considerably from each other. Their forms are listed below in Table 2.

Table 2. Forms of ‘give’ in varieties of Hui Chinese2.

Datapoint of Hui Varieties	Subgroup	Forms of GIVE	Reconstructed Etymon
Xianggao 向杲	Jishe	k ɤ35 or kei35	*kuo1 or *kuo5 ‘give’ 過
Huayang 華陽	Jishe	xɑ̃53	?
Jingzhou 荊州	Jishe	x ɛ35	?
Sanyang 三陽	Jishe	x ɐ33 or x ɑn33	?
Chengkan 呈坎	Jishe	x ɛ42	?
Daguyun 大穀運	Jishe	x ɤ31	?
Tangkou 湯口	Jishe	x ɑ31	?
Huicheng 徽城	Jishe	xe31	?
Hongtan 宏潭	Xiuyi	x ɐ 55	?
Haiyang 海陽	Xiuyi	te55	*diai6 ‘pass’ 遞
Xikou 溪口	Xiuyi	ti55	*diai6 ‘pass’ 遞
Tunxi 屯溪	Xiuyi	ti42	*diai6 ‘pass’ 遞
Liukou 流口	Xiuyi	p ɤ33	*pi3 ‘give’ 畀 or *pi6 ‘have’ 比
Biyang 碧陽	Xiuyi	p ɛi31	*pi3 ‘give’ 畀 or *pi6 ‘have’ 比
Qishan 祁山	Qiwu	fɑ̃11	*fun1 ‘distribute’ 分
Ruokeng 箬坑	Qiwu	fÃ11	*fun1 ‘distribute’ 分
Ziyang 紫陽	Qiwu	tom44	*ton1 ‘hold’ 端
Jiuchengcun 舊城村	Qiwu	to55	*ton1 ‘hold’ 端
Kengtou 坑頭	Qiwu	kA52	*kip7 ‘give’ 給
Qiukou 秋口	Qiwu	k ɑ55	*kip7 ‘give’ 給
Anling 安淩	Jingzhan	x ɛ̃21	?
Jingde 旌德	Jingzhan	pæ213	*pa3 ‘hold’ 把
Zhanda 占大	Jingzhan	p ɔ35	*pa3 ‘hold’ 把
Suian 遂安	Yanzhou	n422	*in4 ‘stretch’ 引
Jiande 建德	Yanzhou	po55	*pa3 ‘hold’ 把
Shouchang 壽昌	Yanzhou	nuə11	*na2 ‘take’ 拿
Chun’an 淳安	Yanzhou	l ɑ53	*na2 ‘take’ 拿

Note: * suggests reconstructed form in historical linguistics.

Table 2. Forms of ‘give’ in varieties of Hui Chinese2.

Datapoint of Hui Varieties	Subgroup	Forms of GIVE	Reconstructed Etymon
Xianggao 向杲	Jishe	k ɤ35 or kei35	*kuo1 or *kuo5 ‘give’ 過
Huayang 華陽	Jishe	xɑ̃53	?
Jingzhou 荊州	Jishe	x ɛ35	?
Sanyang 三陽	Jishe	x ɐ33 or x ɑn33	?
Chengkan 呈坎	Jishe	x ɛ42	?
Daguyun 大穀運	Jishe	x ɤ31	?
Tangkou 湯口	Jishe	x ɑ31	?
Huicheng 徽城	Jishe	xe31	?
Hongtan 宏潭	Xiuyi	x ɐ 55	?
Haiyang 海陽	Xiuyi	te55	*diai6 ‘pass’ 遞
Xikou 溪口	Xiuyi	ti55	*diai6 ‘pass’ 遞
Tunxi 屯溪	Xiuyi	ti42	*diai6 ‘pass’ 遞
Liukou 流口	Xiuyi	p ɤ33	*pi3 ‘give’ 畀 or *pi6 ‘have’ 比
Biyang 碧陽	Xiuyi	p ɛi31	*pi3 ‘give’ 畀 or *pi6 ‘have’ 比
Qishan 祁山	Qiwu	fɑ̃11	*fun1 ‘distribute’ 分
Ruokeng 箬坑	Qiwu	fÃ11	*fun1 ‘distribute’ 分
Ziyang 紫陽	Qiwu	tom44	*ton1 ‘hold’ 端
Jiuchengcun 舊城村	Qiwu	to55	*ton1 ‘hold’ 端
Kengtou 坑頭	Qiwu	kA52	*kip7 ‘give’ 給
Qiukou 秋口	Qiwu	k ɑ55	*kip7 ‘give’ 給
Anling 安淩	Jingzhan	x ɛ̃21	?
Jingde 旌德	Jingzhan	pæ213	*pa3 ‘hold’ 把
Zhanda 占大	Jingzhan	p ɔ35	*pa3 ‘hold’ 把
Suian 遂安	Yanzhou	n422	*in4 ‘stretch’ 引
Jiande 建德	Yanzhou	po55	*pa3 ‘hold’ 把
Shouchang 壽昌	Yanzhou	nuə11	*na2 ‘take’ 拿
Chun’an 淳安	Yanzhou	l ɑ53	*na2 ‘take’ 拿

Note: * suggests reconstructed form in historical linguistics.

When examining the forms of ‘give’ in these Hui varieties, we can identify no fewer than ten distinct forms of the lexical verb ‘give’ in the field (refer to the column ‘Forms of GIVE’ in Table 2), despite the relatively small size of the Hui region (around 25,000 km²) (Zhao 2012), plus the genetic and geographical proximity of these varieties. It is indeed striking because the lexical item GIVE is conventionally used to evaluate the genetic relatedness of languages, which is included not only in the intuition-based Swadesh 100- and 200-word lists (Swadesh 1951, 1955), but also the empirically-based Leipzig-Jakarta list (Tadmor 2009). Therefore, why should these genetically closely related languages of Hui Chinese differ extensively from each other with regard to such a basic lexical item? What is the nature of this variation within the class of these GIVE verbs?

While Hui Chinese lacks a writing system, Hui and Lu (2023) have conducted phonological reconstructions based on the Common Dialectal Chinese (CDC) proposed by Norman (2006) to restore the etymons of these so-alleged GIVE verbs. Based on their findings, we are able to classify these ten different forms of GIVE into three types of etymon verbs (refer to the column ‘Reconstructed Etymon’ in Table 2), which are all closely related to the prototypical ditransitive verb ‘give’, according to Malchukov et al.’s (2010, p. 55) semantic map of ditransitive verbs:

(i)

GIVE-type verb of caused possession, such as *kuo1 or *kuo5 ‘give’ [+give, + manner], *fun1 ‘distribute’ [+give, +manner], *pi6 ‘give’ [+give, -manner] or kip7 ‘give’ [+give, -manner];

(ii)

SEND-type verbs of caused motion in an ‘allative path’ (Malchukov et al. 2010, p. 54), such as *diai6 ‘pass’ [+send, +manner];

(iii)

TAKE-type verbs in an ‘instrumental path’ (Malchukov et al. 2010, p. 54), such as *pa3 ‘hold’ [+take, +manner], *na1 ‘take’ [+take, +manner] and *in4 ‘stretch’ [+take, +manner].

The distinctive etymons of modern Hui varieties play an important role in explaining such radical cases of polygrammaticalization or grammaticalization chains (Craig 1991; Heine et al. 1991) of ‘give’, according to the ‘persistence’ principle of grammaticalization, when some reminiscence of the original forms is likely to retain (Hopper 1991). We will discuss this point further in Section 5.

Having looked at its forms, we will now turn to the functional distribution of the lexical verb ‘give’ in these datapoints.

4. The Polyfunctionality of GIVE in Hui Chinese

One of the features of Sinitic grammaticalization lies in the syncretism of different grammatical functions (Matthews and Yip 2009; Ansaldo et al. 2018). In our investigation, the lexical verb ‘give’ in these Hui varieties also co-exists with all of its grammaticalized forms. Aside from semantic bleaching, phonological reduction is sometimes observed, although it may be more evident in some grammaticalized forms than others in certain Hui varieties.

In sum, we identify a total of ten concomitant functions of the lexical verb GIVE in 27 varieties of Hui Chinese (The Supplementary Material Table S1, namely (i) the lexical verb ‘give’, (ii) recipient marker ‘to’, (iii) benefactive marker ‘for’, (iv) purpose marker, (v) permissive marker, (vi) passive marker, (vii) pretransitive disposal marker, (viii) allative marker, (ix) locative marker ‘at/in’, and (x) temporal marker ‘till’. Among these functions, the last three functions are little discussed in the literature. Given the density of ‘polysemy sharing’ (Chappell and Lü 2022) in a comparatively small geographical area, we cannot help but wonder what the relationships between these distinctive functions are and why some functions are missing in certain Hui varieties. We will start with a descriptive analysis of these functions, before providing explanations.

4.1. Lexical Verb GIVE

The morphemes of GIVE listed in Table 2 can all be used in an unmarked double object construction [SUBJ give OBJ_Rpt OBJ_Th] in Example (1)3. An alternation in a marked indirect ditransitive construction, equivalent to the ‘dative shift’ in English, is attested in most Hui varieties, too, as illustrated in Example (2).

(1)

Shexian (Daguyun) Hui (Chen 2013, p. 155)

x ɣ31	ɑ35	i42	pəŋ35	ɕy31.
give	1SG	one	CLF	book
‘Give me a book.’

(2)

Tunxi Hui

ɑ24	ʨhiu5i ɑ44	ʦʰo11⁻21 ȵie11	ti42- ʨ hio	li ɑu24	l ɔ44	ʨie11- ʦɿ 31
1.SG	uncle	yesterday	give-PFV	two	CLF	chick-egg
‘My uncle (on my mom’s side) gave two baskets of eggs…’
ti42	ɑ24i ɑn44.
RPT	1PL
‘…to me yesterday.’

What is worth pointing out is that while most Hui varieties in our investigation adopt the Standard Mandarin type of unmarked double object construction [SUBJ give OBJ_Rpt OBJ_Th] in Example (1), some Hui varieties, like Shouchang Hui, showcase an opposite word order of the recipient and the theme in Example (3b) like the Cantonese type [SUBJ give OBJ_Th OBJ_Rpt] (Chin 2011; Matthews and Yip 2011), in addition to the type shown in Example (3a). The relative word order of these two variants of the double object construction, i.e., the IO-DO order in ditransitives, is conventionally regarded as a parameter of the North–South division between Sinitic languages (Hashimoto 1976). The fact that Shouchang Hui allows flexible linear word order of the recipient and the theme in Examples (3a) and (3b) suggests possible language contact in its historical development.

(3)

Shouchang Hui (Cao 1997, p. 320–21)

a.	nuə11	ʦ ɑ52	iə ʔ3	pen24	ɕy112!
	give	1SG	one	CLF	book
	‘Give me a book.’
Or
b.	nuə11	iə ʔ3	pen24	ɕy112	ʦ ɑ52.
	give	one	CLF	book	1SG
	‘Give me a book.’

4.2. GIVE as a Recipient Marker

The recipient function of the lexical verb ‘give’ is one of the most commonly attested instances of the semantic shift of GIVE. In our study, the GIVE morpheme in all Hui varieties has a recipient marking function, too, exemplified by the morpheme fɑ̃¹¹ ‘give/RPT’ in Qimen Hui (Example (4)). Besides Sinitic languages, this function is documented in West African languages such as Akan (Example (5)), Ewe, and Yoruba, and Mainland Southeast Asian languages like Thai, Malay, and Vietnamese (Lord et al. 2002).

(4)

Qimen Hui (Hirata 1998, p. 300)

fɑ̃11	n ɔ̃35	k ɔ⁰	m ɑ33	nɑ̃ 11	fɑ̃11	ʂɯ:ə42.
DIS	that	CLF	thing	bring	RPT	1SG
‘Bring that to me.’

(5)

Akan (Lord et al. 2002, p. 219)

me-t ɔ- ɔ	b ɔɔl	no	ma-a	no.
I-throw-PST	Ball	DEF	give-PST	him
‘I threw the ball at him’.

4.3. GIVE as a Benefactive Marker

The benefactive function of GIVE is sometimes discussed together with the recipient function in the literature. Newman (1996) and Lord et al. (2002) account for their relevance in that the recipient can potentially benefit from the act of giving; thus, a ‘benefactive’ reading can sometimes apply to a recipient. Example (6) (Lord et al. 2002, p. 220) demonstrates such a case in Thai, in which the morpheme GIVE can be interpreted both as a recipient and a benefactive marker.

(6)

Thai (Lord et al. 2002, p. 220)

chán	kh ǐan	cotmaăay	hây	kháw	(4)
I	write	letter	give	he
‘I wrote a letter for him/to him.’

Despite a difference in terminology, what we would like to refer to as a ‘benefactive marker’ in this study is not the post-verbal recipient/benefactive marker, but the pre-verbal benefactive marker in the construction [NP₁ BEN NP₂ VP], like yu ‘give’ in Medieval Chinese (Example (7)).

(7)

Medieval Chinese yu ‘BEN’ (Sun 1996, p. 22)

yu	lao	seng	guo	jing	shui-ping
for	old	monk	pass	clean	water-bottle
‘(Someone) rinsed the bottle clean for the old monk.’

This type of pre-verbal benefactive construction is commonly observed in peripheral Hui varieties, i.e., the Jing-zhan and Yanzhou subgroups of Hui, e.g., Jingde Hui, Zhanda Hui, Suian Hui, Jiande Hui, and Chun’an Hui. It is also possible in one central Hui variety, namely Biyang Hui of the Xiyi Group. An example from Jiande Hui is shown below.

(8)

Jiande Hui (Cao 2017, p. 280)

po55	ɑ ŋ213	ɕie55	f ɑom423	ɕin334.
BEN	1SG	write	CLF	letter
‘Write a letter for me.’ (po55 = ‘give’)

4.4. Give as a Purpose Marker

Another function pertinent to the recipient marker is the purpose function. It is sometimes termed a complementizer (Bisang 1996; Lai 2001). Due to the fact that Hui varieties generally lack overt morphological marking for syntactic relationships, we opt to use the term ‘purpose marker’ instead of ‘complementizer’ to emphasize its cognitive schema.

Since the recipient phrase ‘to somebody’ naturally represents the goal of a giving act, which can then be further extended from a person (‘to somebody’) to an event (‘for somebody to do something’), it is not surprising that the same GIVE morpheme for the recipient can also be employed to mark purpose by way of verb serialization, a highly productive strategy in West African languages and Mainland Southeast Asian languages, including Sinitic languages. Although our investigation and the existing literature only offer concrete examples of purpose marking in Tunxi Hui (9) and Wuyuan Hui, we can reasonably infer that this function is likely present in all Hui varieties based on comparative linguistic data from other Sinitic languages like Hakka (Lai 2001) and Cantonese (Chin 2011).

(9)

Tunxi Hui

pu11-21	in 11	lən24	fuə11	ti42	ɑ24	ʨʰi ʔ5,
NEG	need	make	rice	PURP	1SG	eat
‘No need to cook for me.’
ɑ24	ʨiɑu24miɛ11	ʨʰiʔ5	liu24	ʨʰio.
1SG	just now	eat	ANT	SFP
‘I just had something.’

4.5. GIVE as a Permissive Marker

Permission is another widely adopted function of the morpheme GIVE in Hui Chinese. Out of 27 datapoints, the GIVE morphemes in 24 Hui varieties share the identical form with their permissive markers, except for Suian Hui, Chun’an Hui, and Shouchang Hui of the Yanzhou subgroup. An example in Tunxi Hui is demonstrated in Example (10).

(10)

Tunxi Hui

n ̩44	p ɑ 24	pu11	ti42	ɑ 24	kə ʔ5	n ̩44	k ɑ u42.
2SG	dad	NEG	PERM	1SG	COV.with	2SG	talk
‘Your dad doesn’t allow me to tell you.’ (ti42 = ‘give’)

4.6. GIVE as a Passive Marker

A related function of permission in Hui Chinese is the passives. The distribution of passive markers showcases an areal pattern that overlaps with that of the permissive marker mentioned above (Figure 4)5. In other words, a polysemy ‘split’ (Stassen 1997; Koch 2012; Chappell and Lü 2022) is attested in 27 Hui varieties, where the majority (24/27) make use of the same morpheme for GIVE, PERMISSIVE and PASSIVE, whereas among the three exceptions, Chun’an Hui and Suian Hui employ the SUFFER-type passives iə²⁴/tsʰə³¹ ‘meet, suffer’ and fɑ̃⁴²² ‘attack/suffer’, respectively, and Shouchang Hui adopts the CAUSATIVE-type of passives iɑ̃33 ‘causative’.

Examples (11) to (15) demonstrate the GIVE-passive, SUFFER-passive, and CAUSATIVE-passive in four different Hui datapoints.

(11)

Daguyun Hui (GIVE-passive) (Chen 2013, p. 158)

n ɑ33	ʨi42	u ɑ35	xɤ31	kei55
that	CLF	bowl	PASS	3SG
‘That bowl was broken…’
tɑ35-55	pʰo214	tʰɑ42	li.
hit	broken	RVC	SFP
‘…by him/her.’ (xɤ31 = ‘give’)

(12)

Jiande Hui (GIVE-passive) (Cao 2017, p. 330)

mə ʔ2	ʦ a ɑ 55	u ɛ 213	po55	ki334	k ɔ42	pʰ ɑ 213	p ə ʔ 5.
that	CLF	bowl	PASS	3SG	hit	RVC	SFP
‘That bowl was broken by him/her’. (po55 = ‘give’)

(13)

Chun’an Hui (SUFFER-passive) (Cao 2017, p. 330)

len24	k ɑ0	uɑ̃55	tshə31/iə24	kh ɯ445	t ɑ ʔ5	pʰ ɑ24	p ɑ0.
that	CLF	bowl	PASS	3SG	break	RVC	SFP
‘That bowl was broken by him/her’. (tshə31/iə24 = ‘suffer’)

(14)

Suian Hui (SUFFER-passive) (Cao 2017, p. 330)

i33	k ɛ422	uɑ̃213	fɑ̃422	kh ɯ33	t ɑ24	pʰ ɑ55	l ɛ0.
that	CLF	bowl	PASS	3SG	break	RVC	SFP
‘That bowl was broken by him/her’. (fɑ̃422 = ‘suffer’)

(15)

Shouchang Hui (CAUSATIVE-passive) (Cao 2017, p. 330)

mi24	k ɑ33	ŋuə24	iɑ̃33	kə ɯ52	kʰ ɤ11	pʰ ɑ33	p ɑ0.
that	CLF	bowl	PASS	3SG	break	RVC	SFP
‘That bowl was broken by him/her’. (iɑ̃33 = ‘causative’)

The isolated cases of the SUFFER-passives in Chun’an iə²⁴/tsʰə³¹ ‘meet, suffer’ and Suian Hui fɑ̃⁴²² ‘attack/suffer’ are likely attributable to government-initiated migration. According to Cao (2002), government-led migration of the Xin’anjiang Reservoir area in Zhejiang Province has resulted in language enclaves of the original inhabitants, for instance, in Chun’an and Sui’an counties, who used to live near the present-day Xin’anjiang Reservoir and the Thousand Islands Lake and had to emigrate to and scatter in neighboring provinces, Jiangxi in particular. As an exemplification, Cao (1997) postulates that the geographical and eco-cultural factors of the Suian Hui account for such an isolated case, namely mountainous areas, self-contained transportation, and economy, as well as lack of education. The CAUSATIVE-passive in Shouchang Hui, on the other hand, may be due to language contact with Northern Mandarin varieties, the ‘base camp’ of this type of passive.

We will provide further analysis of the source of agent markers in Hui varieties from the perspective of language contact and micro-areal typology in Section 5.

4.7. GIVE as a Pretransitive Disposal Marker

Though the Hui languages under investigation exhibit a canonical SVO word order, the SOV word order is also plausible via secondary topicalization or a ‘pretransitive construction’, in which the overly marked object precedes the main verb (Chao 1968; Matthews and Yip 2008). Pretransitive constructions can be further categorized into the disposal type and the causative type. A much-discussed case in Standard Mandarin is bǎ ‘PRET’, marking either a disposal construction stating how an entity is affected or manipulated (Wang 1947; Li and Thompson 1981), or a causative construction causing an experiencer to experience some emotions6. In this study, by disposal marker, we refer to the overt object marker for the subtype of pretransitive constructions with the semantics of disposal, as compared to those with the semantics of causation. It is also conventionally termed as the ‘object marker’ in Chappell (2015). Like the permissive–passive function, we have also observed a geographical clustering of the distinct semantic sources of pretransitive disposal marker (Figure 5)7.

Three distinct sources of pretransitive disposal markers are found in 27 Hui varieties, namely GIVE-disposals in Example (16), HELP-disposals in Example (17), and HOLD/TAKE-disposals in Example (18). Indigenous Hui groups such as Xiu-yi and Ji-she Hui varieties tend to employ the pretransitive disposal marker grammaticalized from the lexical verb ‘help’, whereas peripheral Hui groups such as Qi-wu and Jing-zhan commonly make use of grammatical markers developed from the lexical verb ‘give’. Sporadic HOLD-type pretransitive markers are also attested in some Hui varieties, especially in the Ji-she and Yanzhou Hui subgroups. Notice that the simultaneous use of pretransitive disposal markers with two or more sources is identified in some varieties, too. For instance, Qishan Hui, Suian Hui, and Chun’an Hui utilize both GIVE and HOLD disposals, and Liukou Hui allows both GIVE and HELP disposals, whereas Sanyang Hui and Huicheng Hui employ both HELP and HOLD disposals. In the extreme case of Haiyang Hui, three lexical sources co-exist as pretransitive disposal markers. Chappell (2015) observes a similar phenomenon in her pan-Sinitic study on object markers and categorized the GIVE and HELP disposals as an areal feature for Central Transitional Sinitic languages, including Hui. Our data further support her observation that the GIVE and HELP disposals may indeed be an innovation unique to Central Transitional and Southeastern Sinitic languages, including some typical Hui languages. The co-occurrence of two or more disposal markers in some varieties, on the other hand, provides another piece of evidence that the HOLD-type disposals in these varieties are likely a ‘borrowed’ form modeled on Standard Mandarin, in competition with the indigenous pretransitive disposal markers grammaticalized from HELP or GIVE verbs, as in Qishan Hui, Suian Hui, Chun’an Hui, Liukou Hui, and Haiyang Hui.

(16)

Fuliang Hui (GIVE-disposal) (Xie 2012, p. 104)

n ̩21	to55	mən24	ko55	t ɑu21.
2SG	DIS	door	close	RVC
‘(You) please close the door!’ (to55 = GIVE)

(17)

Daguyun Hui (HELP-disposal) (Chen 2013, p. 156)

pa31	na33	ko	m ɣ33s ɿ33	t ɑ31	x ɣ31	ɑ35	n ɣ.
DIS	that	CLF	thing	take	RPT	1SG	SFP
‘(Please) take that thing to me.’ (pa31 = HELP)

(18)

Hongtan Hui (HOLD-disposal) (Yuan 2009, p. 117)

n45	n ɐ55	pA55	ʃiE33	kh ɔ̃213	x ʌ31,
2SG	need	DIS	money	put	RVC
‘(You) Need to put away your money,’ (pA55 = HOLD)
p ʌ45	io213	tE45/th ʌ45	tio213	lA0.
NEG	let	fall/take.off	RVC	SFP
‘Don’t let it drop.’

4.8. GIVE as an Allative Marker

The allative function of GIVE indicating directionality is observed in Tunxi Hui in our own fieldwork, although this function is presumed to exist in the two adjacent datapoints of Haiyang Hui and Xikou Hui8 as well, with a shared etymon reconstructed as *diai6 (See Table 2). An example is given in Example (19).

(19)

Tunxi Hui

n ̩44i ɑn44	ʨʰiu11- ȵi ɛ44	ti42	tu ə11-21- ʨʰie11	ɕi11	m ɔ31	ʨi ɑn42	ɕi44,
2PL	last-year	ALL	Tunxi	play	that	CLF	time
‘When you came to Shanghai for sightseeing last year,’
ɑ24	ʨi ɛ42xə31	pu11	ɕi24	k ɔ11	li24.
1SG	right	NEG	COV.at	home	in
‘I happened to be not at home.’

4.9. GIVE as a Locative Marker

Another related function of the allative, i.e., locative function is also evident in Tunxi Hui in Example (20).

(20)

Tunxi Hui

A:	x ɛ̃44-le	ʨy11	ti42	l ɑ ʔ31-5li	ɑ?
	PN.red-DIM	live	LOC	where	Q
	‘Where does Little Red live?’
B:	khə44	ʨy11	ti42	lə ʔ24-5-k ɑ11	mo42li.
	3SG	live	LOC	old-street	there
	‘She lives near the Old Street.’

A bridging context (Evans and Wilkins 2000; Heine 2002) of the allative and locative functions is found in Example (21), which can express either the allative function ‘to’ or the locative function ‘at/in’. However, given a disambiguating context, e.g., whether this conversation suggests a habitual event with some shared knowledge as in Example (22) or a specific event, e.g., heading for school in Example (23), both the speaker and hearer will be able to tell apart the intended function of ti⁴², which shares the same form with the lexical verb ‘give’.

(21)

Tunxi Hui ti⁴² ‘ALL/LOC’

n44	ti42	xo ʔ11-5t ɑu44	tso42	tə ʔ5-mə11?
2SG	ALL/LOC	school	do	what
‘What are you going to school for?’ or ‘What are you doing at school?’

(22)

Tunxi Hui ti⁴² ‘LOC’

A:	n ̩44	ti42	lɑ ʔ24-5li11	tso42	s ɿ11	ɑ?
	2SG	LOC	where	do	thing	Q
	‘Where do you work?’
B:	ɑ24	ti42	xo11t ɑu44	tso42	s ɿ11
	1SG	LOC	school	do	thing
	‘I’m working at school.’
A:	n ̩44	ti42	xo ʔ11-5t ɑu44	tso42	tə ʔ5-mə11	s ɿ11	ɑ?
	2SG	LOC	school	do	what	thing	Q
	‘What do you do at school?’
B:	ko24	ue ʔ11-5 ɕi ɛ11	u ɑ11!
	do	hygiene	SFP
	‘(I’m) cleaning (at school).’ (‘I am a janitor at school.’)

(23)

Tunxi Hui ti⁴² ‘ALL’

A:	n44	ti42	xo ʔ11-5t ɑu44	tso42	tə ʔ5-mə11?
	2SG	ALL	school	do	what
	‘What are you going to school for?’
B:	t ɔ11	mə11-21-s ɿ11	u ɑ11!
	take	thing	SFP
	‘To retrieve something.’

4.10. Give as a Temporal Marker ‘Till’

Like the allative and locative functions, the temporal marking function of the morpheme GIVE is only attested in Tunxi Hui, as shown in Example (24).

(24)

Tunxi Hui

ti42	m ɛ44-ni ɛ44	ʦ ə31	k ɑu31	u ɛ31 ǃ
ALL	next-year	again	talk	SFP
‘(Let’s) talk about it next year!’

Another morpheme tɛ³¹ ‘wait’ is competing with ti⁴² ‘pass/give’ for the same function of temporal marking for future events, as shown in Example (25).

(25)

Tunxi Hui

t ɛ31	m ɛ44-ni ɛ44	ʦ ə31	k ɑu31	u ɛ31 ǃ
wait	next-year	again	talk	SFP
‘(Let’s) talk about it next year!’

Although both constructions are found in natural recordings, there is an observed tendency for speakers to prefer tɛ³¹ ‘wait’ to ti⁴² ‘pass/give’ in marking this temporal relationship.

For detailed analyses of the allative, locative, and temporal functions of GIVE in Tunxi Hui, please refer to Lu and Hui (2023).

4.11. Give as an Imperative Marker

The last co-existing function of GIVE involves six datapoints in our investigation. When it occurs, it is largely accompanied by the first-person singular pronoun ‘1.SG’, i.e., ‘give-me’, normally with the subject being the second-person pronoun, as in Examples (26) and (27).

(26)

Xianggao Hui (Shen 2012, p. 123)

(n ̩ 35)	(kɤ35-ŋ ɔ35)	ʨʰi32	k ɒ52 ʦʰin22	n ̩ 22	u ɒ35	f ɒ22 ǃ
2SG	IMP-1SG	eat	clean	this	CLF	rice
‘(You) please finish this bowl of rice!’

(27)

Jixi Hui (Zhao 2003, p. 139)

(n213)	(po213- ɔ213)	ku ɑ213	tsh ɣʔ32	khi35!
2SG	IMP-1SG	roll	out	go
‘(You) get out of here!’

The imperative use of GIVE is not a novel invention of Hui Chinese. As a matter of fact, it is also present in Mandarin varieties (28), as well as genetically unrelated languages like Russian (29) and Italian (Kuteva et al. 2019).

(28)

Xuzhou varieties of Central Plains Mandarin (Kuteva et al. 2019, p. 198)

ke55	uo213	i213t ɕi35t ɕʰy51.
IMP	1SG	stand.up
‘Stand up!’

(29)

Russian (Kuteva et al. 2019, p. 197)

Davaj	pojdjom	v	kino!	(9)
give.IMP.SG	go.1PL	in	movies
‘Let’s go to the movies!’

In this section, we have provided a descriptive analysis of a multitude of functions associated with the morpheme GIVE, as well as their distributions in 27 varieties of Hui Chinese. In the next section, we aim to account for such radical polyfunctionality from the perspective of semantic extension, polygrammaticalization, and cooptation.

5. Explanation for Polysemy Sharing of GIVE

Semantic extension, polygrammaticalization, and cooptation join hands in shaping the multiplicity of functions found in the Hui varieties in our study. We will address these mechanisms one by one.

5.1. Semantic Extension

The distinctive forms of the morphemes for ‘give’ are a notable characteristic in Hui varieties of Sinitic languages (Table 2). It is unusual to observe such a significant divergence in closely related and geographically adjacent language varieties for a presumably basic concept ‘give’. As an explanation, Zhang (2011) suggests that the ‘give’ verbs in modern Sinitic languages may not have originated as the prototypical ditransitive verb ‘give’, but as verbs meaning ‘send’, ‘take’, ‘hold’, etc. Fortunately, phonological reconstructions of relevant verbs in Hui varieties (Hui and Lu 2023) shed light on their etymons, confirming their origin as verbs for SEND, TAKE, and HOLD. These verbs have likely extended to ‘give’ both semantically and functionally from a peripheral ‘give’ verb to the prototypical ditransitive ‘give’ verb, based on a route proposed by Malchukov et al. (2010, p. 55) on ditransitive verbs.

Parallel to the difference in forms, there are three functions of GIVE (namely allative, locative, temporal) that are only observed in Tunxi Hui, and possibly in its neighboring varieties of Hui Chinese with cognates of ti⁴². These functions, however, are explicable only if they were grammaticalized not from GIVE, but from a SEND-type verb ‘pass’, the etymon of which Tunxi Hui ti⁴², Xikou Hui ti⁵⁵, and Haiyang Hui te⁵⁵ can be traced back to. This is because directionality and the sense of ‘goal’ are inherently embedded in the semantics of ti⁴² ‘pass’. By extending the concept of goal from a human recipient to a location, PASS grammaticalizes into an allative marker. From the allative [+location, +direction] function, a locative function of ti⁴² ‘at/in’ is developed by losing its semantic feature [+direction]. Lastly, the temporal marking of ti⁴² ‘to, till’ can be explained with the ‘ubiquity of conceptual transfer from time to space’ (Haspelmath 1997, p. 140) attested cross-linguistically, as ti⁴² ‘to, towards’, an allative in location, can be extended to ti⁴² ‘to, till’, an allative in time. A detailed analysis of this grammaticalization chain can be found in Lu and Hui (2023).

To summarize, the lexical verb corresponding to the sense of GIVE may have been missing in Hui Chinese. Instead, SEND-type verbs like ti⁴², te⁵⁵, and ti⁵⁵ with the etymon *diai6 ‘pass’ [+send, +manner], and TAKE-type verbs like tom⁴⁴, to⁵⁵, n⁴²², nuə¹¹, and la⁵³ with the etymons *ton1 ‘hold’, *pa3 ‘hold’, *in4 ‘stretch’, and *na2 ‘take’ may have emerged as winners in the competition against other verbs with similar semantics. Through semantic extension, these verbs gradually took over both the meaning and the function of the prototypical ditransitive verb ‘give’ in each of the Hui varieties.

Following this semantic extension from SEND or TAKE to GIVE, separate grammaticalization pathways emerge, including one grammaticalization cline starting from SEND and leading to the development of an allative marker before further evolving into locative or temporal markers, and the multiple grammaticalization clines in relation to GIVE. In the next section, we will discuss the mechanism of these grammaticalization pathways in greater detail.

5.2. Polygrammaticalization

Besides semantic extension, grammaticalization can account for the majority of the multiple functions of GIVE in our study of Hui varieties. There are five separate grammaticalization clines that can be identified among the ten functions of ‘give’ in this study:

• Lexical verb ‘give’ > Benefactive ‘for’
• Lexical verb ‘give’ > Recipient ‘to’
• Lexical verb ‘give’ > Recipient ‘to’ > Purpose marker
• Lexical verb ‘give’ > Recipient ‘to’ > Pretransitive disposal marker
• Lexical verb ‘give’ > Causative-Permissive > Passive

To begin with, the lexical verb ‘give’ can also function as a benefactive marker, as discussed in Newman’s (1996) cognitive analysis of ‘give’. This function becomes apparent when ‘give’ collocates with ‘donary verbs’, as the prototypical sense of giving inherently carries a nuance of benefiting (Newman 1996, pp. 218–19). This grammaticalization pathway is also reported in the World Lexicon of Grammaticalization (Kuteva et al. 2019).

The next grammaticalization cline (ii) from the lexical verb ‘give’ to the recipient ‘to’ is also widely attested cross-linguistically. It is not only a pan-Sinitic phenomenon, but is also well-established in languages of West Africa and East and Southeast Asia (Lord et al. 2002). During this process, the semantics of the lexical verb ‘give’ are bleached. Hence, the lexical verb ‘give’ is decategorized into a functional category, i.e., a ‘coverb’ (Matthews and Yip 2011) or preposition by reanalysis via serial verb constructions [give NPth RPT marker NPrpt].

From the recipient marker, the morpheme ‘give’ then develops two new functions, (iii) the purpose marker and (iv) the pretransitive disposal marker. Although both are outcomes of grammaticalization, they differ in each other in that the purpose marker cline is a case of cognitive-driven language-internal grammaticalization, whereas the pretransitive disposal marker pathway is an instantiation of contact-induced replica grammaticalization (Heine and Kuteva 2003, 2005; Matthews and Yip 2009). Such a difference in grammaticalization mechanisms also results in a divergence in the distribution of these two functions. The purpose marker function is fairly common in many Sinitic languages, by extending the goal of the action of giving from a recipient ‘RPT’ [+ human being] to an event ‘PURP’ (Lai 2001; Lu and Hui 2023). The pretranstive disposal marker function of ‘give’, however, is scarcely recorded in the literature, except for a few works, e.g., on Central Transitional Sinitic languages in Chappell’s (2015) investigation as well as this study. Besides, this lexical–functional sharing displays geographical clustering not only within Hui varieties (Figure 5), but also in Chappell’s (2015) pan-Sinitic study on disposal markers as an area-specific feature of Central Transitional Sinitic languages. Its scarcity and areal convergence point to the likelihood of a type of contact-induced grammaticalization known as replica grammaticalization, in which the grammaticalization process that has taken place in the model language is transferred to the replica language (Heine and Kuteva 2003, 2005). What makes the transfer of this grammaticalization process from ‘give’ to pretransitive disposal marker possible is probably the underlying etymology of GIVE in the replica language. Many varieties of Hui displaying the ‘give’-disposal grammaticalization cline happen to possess a GIVE verb that originated as TAKE or HOLD verbs (see

Table 3. GIVE as pretransitive disposal markers in Hui varieties.

Datapoint	Subgroup	GIVE	Reconstructed Etymon	Types of Verbs	GIVE as DISP
Haiyang (Xiuning)	Xiuyi	te55	*diai6 ‘pass’ 遞	PASS	+
Liukou (Xiuing)	Xiuyi	p ɤ33	*pi3 比or *pi6 ‘give’畀	GIVE	+
Biyang (Yixian)	Xiuyi	p ɛi31	*pi3 比or *pi6 ‘give’畀	GIVE	+
Qishan (Qimen)	Qiwu	fɑ̃11	*fun1 ‘distribute’ 分	GIVE	+
Fuliang	Qiwu	to55	*ton1 ‘hold’ 端	TAKE/HOLD	+
Jingde	Jingzhan	pæ213	*pa3 ‘hold’ 把	TAKE/HOLD	+
Zhanda	Jingzhan	p ɔ35	*pa3 ‘hold’ 把	TAKE/HOLD	+
Suian (Yanzhou)	Yanzhou	n422	*in4 ‘stretch’ 引	TAKE/HOLD	+
Jiande (Yanzhou)	Yanzhou	po55	*pa3 ‘hold’ 把	TAKE/HOLD	+
Chun’an (Yanzhou)	Yanzhou	l ɑ53	*na2 ‘take’ 拿	TAKE/HOLD	+

). TAKE and HOLD verbs are natural sources for disposal markers in Sinitic languages, e.g., bǎ 把 ‘hold’ in Standard Mandarin. In the case of Hui Chinese, the reminiscence of the etymological meaning of HOLD or TAKE of the current GIVE verbs facilitates the polysemy sharing between ‘give’ and the disposal marker. At first, grammaticalization from HOLD or TAKE to disposal marker took place via the bridging context [Vtake/DISP marker NP_th Vgive NP_rpt]. Later on, after these verbs of HOLD and TAKE were extended to function as the prototypical ditransitive GIVE, adjacent languages likely borrowed this grammaticalization pathway via replica grammaticalization with the synchronic lexical–functional parallel of ‘give’ and the disposal marker. To this end, semantic extension, language-internal grammaticalization, and replica grammaticalization join hands in shaping the concurrent use of ‘give’ and the pretransitive disposal marker in varieties of Hui Chinese.

In addition to the grammaticalization pathway from ‘give’ to disposal marking, another noteworthy path (v) demonstrating areal convergence is the one leading from the lexical verb ‘give’ through permissive to the passive marker10. There is a comparatively rich body of literature on this contact-induced grammaticalization process in specific geographical regions, including Sinitic languages spoken south of the Yangtze River, encompassing Central Transitional, Southeastern, and Far Southern Sinitic varieties like Hui, Wu, Min, Yue, Hakka, and Gan (Lai 2001; Matthews et al. 2005; Chen 2009; Chin 2011; Chappell 2015; Ngai 2015). This process is also observed in some Mainland Southeast Asian languages such as Tai and Austroasiatic languages, along with colloquial Malay (Yap and Iwasaki 2007). The initial transition from the lexical verb ‘give’ to causative/permissive can be explained by the inherent notion of causation within the semantics of GIVE as ‘caused possession’ (Jackendoff 1990). The next stage from causative to passive becomes possible through the concept of ‘unwilling permissives’ (Chang 2006, p. 139) or ‘reflexive’ causatives (Haspelmath 1990, p. 48; Yap and Iwasaki 2007, p. 200), where an agent is made to act unwillingly, resulting in a passive interpretation. Features unique to the ‘give’-passives include the obligatory requirement of the agent phrase and the adversative reading of the event. While the compulsory presence of the agent phrase of ‘give’-passives results from the syntactic configuration of ‘give’ as a three-argument verb, the adversity embedded in ‘give’-passives can be attributed to the intermediate stage of the ‘involuntary’ permissive connotation. Both features demonstrate the principle of ‘persistence’ in grammaticalization (Hopper 1991), in which remnants of the original lexical form, both semantically and syntactically, are preserved in the grammaticalized form.

5.3. Cooptation

The final function of ‘give’ which our study identifies is imperative marking. Unlike the functions discussed above, this imperative use of ‘GIVE-1SG’ represents a discourse marker, the mechanism of which can be explained by cooptation, a process related to but distinct from grammaticalization (Heine 2013; Heine et al. 2017, among others). The operation of cooptation occurs when ‘information units such as clauses, phrases, or words are transferred from the domain of sentence grammar to that of discourse organization’ (Heine 2013, p. 1205). According to Heine (2013), cooptation is characterized by a series of features that set it apart from grammaticalization, e.g., spontaneity, increased scope, syntactic independence from the sentence structure it appears in, metacommunicative meaning, and unrestrictive word order. Many instances of the discourse marker ‘GIVE-1SG’ in Hui Chinese exhibit most of these features.

To begin with, the instantiations of ‘GIVE-me’ occur mostly in spontaneous data, i.e., common in colloquial language instead of written texts. Apart from Hui varieties, this imperative marker is also witnessed in dialogue recordings of many Mandarin varieties, e.g., Jinan, Liuzhou, Nanjing, Xi’an, and Wanrong. An example is shown in Example (30).

(30)

Nanjing Mandarin (Li 2002, p. 4643)

ku ɑe31	(ki11-12-uo11)	tʂʰ ɿ ʔ5	ɕi ɑ44	kʰi 44	(11)
fast	give-1SG	eat	down	go
‘(You) Eat it, quick!’

Besides, this discourse marker is independent of the syntagmatic configuration of the sentence it occurs in. In other words, deleting the discourse marker will not affect the grammaticality of the sentence. For instance, without the phrase kɤ³⁵-ŋɔ³⁵ ‘GIVE-1SG’ in Example (26) of Xianggao Hui or po²¹³-ɔ²¹³ ‘GIVE-1SG’ in Example (27) of Jixi Hui above, the original sentences remain intact.

In addition, the pragmaticalized ‘GIVE-1SG’ focuses on the speaker–hearer relationship, as it is usually recruited in an imperative sentence with an explicitly expressed or sometimes omitted second-person subject ‘you’, i.e., the hearer, instead of the first- or third-person forms. The inclusion of the first-person singular morpheme into the formation of the discourse marker itself in ‘GIVE-1SG’, on the other hand, is suggestive of the illocution of the speaker–hearer relationship as a command ‘for me’. In contrast, a related discourse marker kei³³-ʨʰi^a24 ‘GIVE-1PL.INCL’ found in Wanrong Mandarin, which fuses the first person plural morpheme into the formation of the marker, denotes negotiation ‘for us’. Compare:

(31)

Wanrong Mandarin (Li 2002, p. 4643)

ŋ ɣ55	kei33	ʨʰi ɑ24	ʦ əu33	fæ̃33,	(12)
1SG	BEN	me	do	rice
‘I will do the cooking for us,’
ȵi55	(kei33- ʨʰ ia24)	p ɑ51	ŋ ɑu55	ɕi55	l ɑu!
2SG	GIVE-1PL	DISP	coat	wash	SFP
‘Could you please do the laundry?’

Lastly, the discourse marker ‘GIVE-1SG’ allows relatively free word order. Although ‘GIVE-1SG’ appears pre-verbally at large, it is also acceptable when placed post-verbally at the end of the sentence, albeit far less commonly, as demonstrated in Example (32b) below:

(32)

Colloquial Putonghua

a.	(n ǐ)	(g ěi-w ǒ)	g ǔn	ch ū	q ù!
	2SG	give-1SG	roll	out	go
	‘(You) get out of here!’
b.	(n ǐ)	g ǔ n	ch ū-q ù	(g ěi-w ǒ)!
	2SG	roll	out-go	give-1SG
	‘(You) get out of here!’

In summary, cooptation is the mechanism responsible for the use of the discourse marker ‘GIVE-1SG’, which serves to issue a command for the hearer to do something for the speaker. This marker does not participate in the constituent structure of the sentence, thus allowing for a relatively flexible word order.

5.4. A Summary of the Polyfunctionality of GIVE in Hui Chinese

In this section, we have offered an explanation for all the functions of ‘give’ observed in our study of the 27 varieties of Hui Chinese. The relationships among these multiple functions are schematically represented below in Figure 6:

6. Areal and Typological Implications

Analysis of the polyfunctionality and polygrammaticalization of ‘give’ in Hui Chinese not only deepens our understanding of this under-investigated group of Sinitic languages but also makes a valuable contribution to Chinese linguistics as a whole. Before wrapping up this study, we would like to highlight a couple of topics that warrant further research.

6.1. Central China as a Grammaticalization Area

Analyzing the disposal, passive and comparative constructions across the Sinitic branch, Chappell (2015) classifies Sinitic into five principal ‘linguistic areas’, where Hui (alongside Wu, Gan, Xiang, Jianghuai Mandarin, some subgroups of Southwest Mandarin and enclaves of Hakka) belongs to the Central Transitional areal group. Before we delve further, it is worth discussing Chappell’s (2015) classification scheme in more detail. First, Chappell uses the term ‘linguistic areas’ to refer to the areal groups identified in the study. Given that only Sinitic varieties are taken into account, the use of this term deviates from its generally accepted meaning within areal linguistics, where languages of different families (or at least of distinct branches within a family) have to be involved13 (Campbell 1985; Thomason 2001; Stolz 2002). Second, the proposed areal groups are established based on the lexical sources of the disposal markers, passive markers, and comparative markers of various Sinitic varieties; although the structural type of the comparative construction is also considered, it is strongly interrelated with the lexical sources thereof. In other words, grammaticalization pathways are effectively the sole criteria for determining the areal groups. While grammaticalization pathways are often taken into account when defining linguistic areas14, all firmly established linguistic areas around the globe share features from various domains of grammar (Thomason 2001; Campbell 2020), and those determined exclusively by grammaticalization patterns are sometimes referred to as ‘grammaticalization areas’ instead (Heine 1994; Heine and Kuteva 2005; Huang and Wu 2018). Regardless of the terminological difference, the present study can contribute to our understanding of the convergence phenomena in the Central China Transitional region.

Given their geographical location, it comes as no surprise that Sinitic varieties in Central China feature a mix of Northern and Southern traits. However, what makes this areal group of scholarly interest is the widespread occurrence of grammaticalization patterns which are rare or even unattested in Sinitic varieties spoken elsewhere. As Chappell (2015) observes, such area-specific patterns include the development of ‘give’ and ‘help’ into disposal markers, as well as ‘take’ and ‘wait’ into passive markers. In this study, we demonstrate that the polygrammaticalization of ‘give’ in Hui Chinese is related to its multiple sources of ‘give’ verbs (Table 2). Since this polygrammaticalization phenomenon consists of a series of cognitively driven language-internal grammaticalization pathways (Section 5), a possible explanation for some of the area-specific grammaticalization patterns (e.g., ‘take’ passives) in Central China might be that they originated from Hui Chinese and spread to other Sinitic varieties in the region via replica grammaticalization, where the grammaticalization processes which occurred in Hui Chinese were transferred to other Sinitic varieties in Central China15. Nonetheless, although the geographical location of Hui Chinese makes it a plausible source of these grammaticalization patterns, in order to provide stronger support for this hypothesis, we need to investigate the functions and etymons of ‘give’ in other Sinitic languages in the Central China Transitional area.

6.2. Multiple Lexical Sources of Pretransitive Disposal Markers

As we mentioned in Section 4.7, Hui Chinese varieties commonly employ two pretransitive disposal markers of different lexical sources, with Haiyang Hui representing the most extreme case, where three different disposal markers are used. While this is a rare phenomenon in Sinitic languages as a whole, the co-existence of three or more disposal markers is pervasive in the Chaoshan subgroup of Southern Min, whose division of labor appears to be conditioned by certain semantic and pragmatic factors (Lin and Szeto 2023). Given the multiple lexical sources of disposal markers in Hui Chinese, it holds promise to investigate the functional range of these markers and analyze whether it is affected by their lexical sources (and hence grammaticalization pathways). This research direction can shed new light on the potential effect of a disposal marker’s lexical source on its functional scope, e.g., whether it can pair up only with resultative verbs (e.g., ‘to break’) or also with other transitive verbs that are placed lower on Tsunoda’s (1985) Affectedness Hierarchy (e.g., ‘to hit’, ‘to see’). This remains an important yet inadequately addressed issue in the study of Chinese disposal constructions.

6.3. Diachronic Stability of ‘Give’

The diverse origins of ‘give’ in different varieties of Hui Chinese raise questions about the suitability of this seemingly ‘basic’ verb for assessing genetic relatedness between languages16. The semantic links between ‘give’ and a wide array of concepts (Malchukov et al. 2010; Rzymski et al. 2020) also suggest that it may be prone to semantic shifts and lexical replacements. Interestingly, though, when compiling the Leipzig-Jakarta list based on data from 41 languages around the world, Tadmor (2009) finds that ‘give’ is not only resistant to borrowing but also remarkably stable over time. In fact, ‘give’ appears to be a highly stable lexical item across many well-studied language branches. For instance, the ‘give’ verbs in Germanic languages (e.g., English: give, Dutch: gevan, German: geben, Danish: give, Swedish: giva, Icelandic: gefa) all trace their roots back to Proto-Germanic *gebaną; likewise, those of Tai languages (e.g., Thai: hâi, Lao: hai5, Tai Lü: hɯ3, Tai Nüa: haɯ3, Nùng: hùː, Zhuang: haw3, Bouyei: haec) are uniformly descended from Proto-Tai *haɰ^C. With over 20 distinct etymons of ‘give’ in its language varieties (Cao 2008, L151), Sinitic appears to represent a theoretically reasonable yet cross-linguistically rare case in which ‘give’ has undergone numerous series of lexical replacements throughout its history. The presence of a broad range of GIVE, SEND, TAKE, and HOLD verbs in Old Chinese17 may have contributed to such developmental patterns (see Section 5.1); further studies on the functions and sources of GIVE in other groups of Sinitic can help account for this intriguing phenomenon. What gave rise to the rich diversity of GIVE-related verbs in the first place calls for another line of future research, possibly necessitating cross-disciplinary collaboration.

6.4. Conclusions

This study has provided an areal and typological analysis of the multiple forms and functions of GIVE in 27 datapoints of Hui Chinese, a little-studied group of Sinitic languages. The Hui Chinese varieties examined in this study generally demonstrate the syncretism of at least five functions associated with the morpheme GIVE synchronically. Semantic extension, polygrammaticalization, and cooptation are shown to be the major mechanisms behind the polyfunctionality or polysemy sharing of the morpheme GIVE. Our study contributes to our understanding of the role that grammaticalization, particularly contact-induced grammaticalization, plays in the formation of linguistic areas. Furthermore, it raises questions regarding the suitability of ‘give’ for assessing the genetic relatedness of languages worldwide.