A typology and geolinguistics of sibling term systems in Tibeto-Burman

『地理言語学研究』 | Studies in Geolinguistics 5, 2025 38 Research article A typology and geolinguistics of sibling term systems in Tibeto-Burman KURABE, Keita ILCAA, Tokyo University of Foreign Studies SUZUKI, Hiroyuki Kyoto University SHIRAI, Satoko The University of Tokyo EBIHARA, Shiho Kyoto University IWASA, Kazue Nagoya University of Foreign Studies Abstract: Kinship terminology is a foundational domain in anthropology, providing insights into how languages encode social relations. This study examines sibling term systems in Tibeto-Burman languages from typological and geolinguistic perspectives, identifying eight types defined by relative age, referent’s sex, and relative sex. The typological diversity of these systems mirrors broader patterns of lexico-semantic diversity across Tibeto-Burman languages. While types such as the Undifferentiated Sibling Type, the Relative Sex/Age Type, and the Relative Sex/Sex Type may be considered secondary because of their skewed genealogical and geographical distributions, other types, particularly the Skewed Age Type, the Age/Sex Type, and the Relative Sex/Age/Sex Type, are likely to reflect older strata. These findings contribute to a deeper typological understanding of kinship terminology and its areal distribution in Tibeto-Burman languages. Keywords: Tibeto-Burman; kinship terminology; sibling typology; geolinguistics; language contact 1. Introduction Kinship terminology has long been a central focus in anthropological research, as it provides a window into how languages encode social relations and how such encoding KURABE, Keita, Hiroyuki SUZUKI, Satoko SHIRAI, Shiho EBIHARA, and Kazue IWASA. 2025. A typology and geolinguistics of sibling term systems in Tibeto-Burman. Studies in Geolinguistics 5: 38–65. doi: https://doi.org/10.5281/zenodo.17479479  This work was supported by the ILCAA Joint Research Project “Studies in Asian and African Geolinguistics” (jrp000256) and JSPS KAKENHI Grant Numbers JP25770167, JP16H02722, JP17H04774, JP18H00670, JP23K00476, JP24K03887, JP24K23937, JP25K04040. Any remaining errors are the sole responsibility of the authors. KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 39 reflects underlying cultural structures (e.g., Morgan 1871; Rivers 1914; Murdock 1949). Within this domain, sibling terminology in particular has received sustained attention across disciplines, from Morgan’s pioneering comparative study (1871) to subsequent anthropological and linguistic investigations (Nerlove & Romney 1967; Murdock 1968, 1970; Kronenfeld 1974; Matsumoto 2000; Fukushima 2023, 2024), highlighting its significance in the analysis of kinship systems. Fukushima et al. (2023, Chapter XIV) present a large-scale geolinguistic study of sibling terminology across the languages of Asia and Africa. The study brings together specialists in individual language families or areal groups, including ChukotkoKamchatkan, Ainu, Japonic, Korean, Sinitic, Hmong-Mien, Kra-Dai, Tibeto-Burman, Austroasiatic, Austronesian, Tungusic, Uralic, Mongolic, Turkic, South Asian, Dravidian, Iranian, Semitic, Nilo-Saharan, Niger-Congo, and the Kalahari Basin Area. While the study offers a broad typological and geolinguistic overview, each chapter is kept brief due to space limitations. Building on the study of sibling term systems in Tibeto-Burman presented in the volume (Kurabe et al. 2023), this paper seeks to supplement those findings by providing more detailed data and by extending the discussion in light of subsequent studies and additional relevant literature. This contributes to a broader typological understanding of kinship terminology and its distributional patterns across the Tibeto-Burman languages. The structure of this study is as follows. Section 2 outlines the cross-linguistic typology of sibling term systems. Section 3 details the types attested in Tibeto-Burman. Section 4 examines their geographical distribution and considers possible diachronic developments. Section 5 concludes the paper. 2. Typology of sibling term systems This section provides an overview of the typology proposed in Fukushima (2023), which forms the basis for the description of sibling term systems in Tibeto-Burman in Section 3. The classification builds on earlier work by Murdock (1968) and Matsumoto (2000), and describes systems of sibling terms using three criteria: (1) relative age, (2) referent sex, and (3) relative sex. 2.1. Undifferentiated Sibling Type Languages that fall under the Undifferentiated Sibling Type (Type A in Fukushima 2023) employ a single lexical item to refer to all siblings, without distinctions of age or gender. For example, Ki-Nubi, an Arabic-based creole spoken in Uganda and Kenya, follows this pattern, as the term akú is used without distinction of age or gender. This would schematically be represented as (1) below, where akú uniformly fills the cells 『地理言語学研究』 | Studies in Geolinguistics 5 40 for all combinations of age and gender. When differentiation is required, it is achieved by adding lexical items denoting ‘man’ or ‘woman’ (Nagato 2023). (1) Ki-Nubi (adapted, ibid., p.51) Male Female Elder akú Younger Cross-linguistically, this type is geographically and genealogically limited, occurring primarily in coastal West African languages and in the Philippine languages of Austronesian (Matsumoto 2000: 5). Within Southeast Asia, it is rarely found on the mainland. It is absent in Kra-Dai and Hmong-Mien, and extremely limited in Austroasiatic and Tibeto-Burman (Fukushima et al. 2023). In contrast, it is relatively widespread in the insular region, particularly in Austronesian languages spoken in the Philippines, Taiwan, Indonesia, and across the Pacific (Utsumi 2023). Matsumoto (2000: 6) argues that sibling term systems based on referent age and referent sex are structurally incompatible, so that a shift from one to the other would entail the collapse of the existing system. The Undifferentiated Sibling Type, in this view, arises when a previously established system collapses, allowing underlying neutral or generic terms to emerge as surface forms. 2.2. Relative Age Type This type (Type B in Fukushima 2023) encodes relative age through separate terms for ‘elder’ and ‘younger’ siblings, without reference to the referent’s sex. This is illustrated by Ts’ixa, a language of the Kalahari-Khoe branch within the Khoe-Kwadi family of the Kalahari Basin Area, where lexical distinctions are made based on relative age but not sex. The term táá-xū refers to an elder sibling, while dámā-xù denotes a younger sibling (Kimura & Nakagawa 2023). (2) Ts’ixa (adapted, ibid., p.58) Male Female Elder táá-xū Younger dámā-xù The Relative Age Type is frequently observed across several language families in Southeast Asia, including major languages such as Thai, Khmer, and Indonesian. Within Tibeto-Burman, it is especially prevalent in the Karenic languages (§3.2). Among Tai languages, this type is widely attested, with terms for ‘elder sibling’ and KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 41 ‘younger sibling’ typically reflecting Proto-Tai *biB2 and *nɔŋC2, respectively (Hirano et al. 2023). In Austroasiatic, it is one of the most common patterns, occurring across various subgroups such as Aslian, Bahnaric, Khmeric, Pearic, and Munda (Shimizu & Minegishi 2023). It is also widely distributed across Austronesian languages, particularly in Taiwan, the Philippines, and Indonesia (Utsumi 2023). Outside Southeast Asia, the Relative Age Type is found in the Kalahari Basin Area, and to a lesser extent in the Niger-Congo, Nilo-Saharan, Turkic, Japonic, and South Asian groups. It tends to occur in areas adjacent to those of the Undifferentiated Sibling Type, and its distribution is comparatively denser (Fukushima 2023, 2024). 2.3. Skewed Age Type This type (Type C in Fukushima 2023) refers to the system that distinguishes sex in either elder or younger siblings, but not both. Typically, only elder siblings are differentiated by sex, while the category of younger siblings remains undifferentiated. For instance, a language may have separate terms for ‘elder brother’ and ‘elder sister’ but use a single term for both younger siblings. This pattern is illustrated by Nenets, a Uralic language spoken in northern Siberia, where nyaka refers to an elder brother and nyabako to an elder sister, while papa denotes a younger sibling regardless of sex. When further specification is needed, the referent’s sex is expressed through compounding (Matsumoto 2023). (3) Nenets (adapted, ibid., p.39) Male Female Elder nyaka nyabako Younger papa The Skewed Age Type is widely attested across Eurasia and Africa, with notable concentrations in Southeast and northern Asia. Beyond Afro-Eurasia, it is commonly found in the indigenous languages of Australia (Matsumoto 2000: 8–10; Fukushima 2023, 2024). In Southeast Asia, this type is well represented in Kra-Dai and Austroasiatic, and also occurs with some frequency in Tibeto-Burman and Austronesian languages. Within Kra-Dai, it is primarily found in Northern Tai, particularly in northern Zhuang dialects, as well as in Central Tai, the Li languages of Hainan Island, and Laji in northwestern Vietnam (Hirano et al. 2023). In Austroasiatic, this type, along with the Relative Age Type (§2.2), ranks among the most widespread patterns, being attested across multiple branches including Bahnaric, Katuic, Khmuic, Mangic, Monic, Palaungic, Vietic, and Munda (Shimizu & Minegishi 2023). In Austronesian, it is found in languages spoken in Sumatra and Java (Utsumi 2023). 『地理言語学研究』 | Studies in Geolinguistics 5 42 2.4. Age/Sex Type This type (Type D in Fukushima 2023) features a fully differentiated system, with four distinct terms that encode both relative age and gender: ‘elder brother,’ ‘elder sister,’ ‘younger brother,’ and ‘younger sister.’ An example is provided by Nanjing Mandarin, in which all four categories are lexically distinguished (Yagi 2023). (4) Nanjing Mandarin (adapted, ibid., p.21) Male Female Elder 哥哥 姐姐 Younger 弟弟 妹妹 The Age/Sex Type is one of the most dominant and structurally stable sibling term systems cross-linguistically, both in terms of internal structure and geographical distribution. Within Asia, it is the predominant pattern in Sinitic and is also strongly represented in Japonic, Tibeto-Burman, and Dravidian (Matsumoto 2000: 10–11; Fukushima 2023, 2024). Among the 612 Chinese dialects surveyed by Yagi (2023), only this type was attested. In Southeast Asia, it is particularly prominent in HmongMien and Tibeto-Burman, and is also found in Kra-Dai and Austroasiatic. Within Hmong-Mien, it is widely observed in the eastern region, irrespective of lect affiliation. Its occurrence in Sinitic may reflect the influence of language contact (Taguchi & Tang 2023). In Kra-Dai, the type is well represented in the Kra branch, including languages such as Gelao and Buyang, and is also found in some Northern Tai languages (Hirano et al. 2023), which suggests another instance of contact-induced influence from Sinitic. In Austroasiatic, it is found in the Munda branch, where it constitutes the predominant sibling term system (Shimizu and Minegishi 2023). By contrast, this type is entirely absent from the Austronesian languages surveyed by Utsumi (2023). 2.5. Sex Type The Sex type, also known as the brother/sister type (Type E in Fukushima), distinguishes between ‘brother’ and ‘sister’ regardless of relative age. This pattern is exemplified by Azeri, a Turkic language spoken in western Asia. In Azeri, sibling terms encode the referent’s sex but do not indicate age differences, with kardaš denoting a brother and baǰi a sister (Saitô 2023). (5) Azeri (adapted, ibid. p.41) Male Female Elder kardaš baǰi Younger KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 43 The Sex Type, also referred to as the European or Brother-Sister Type by Murdock (1968), is the most prevalent sibling term system in European languages. This type is strongly associated with the Indo-European and Afro-Asiatic language families and is geographically widespread across northern Africa, most of Europe, and through Southwest Asia into the Indian subcontinent. Although it also appears sporadically in other language families, such associations are likely to result from secondary developments through language contact (Matsumoto 2000: 11–15). In Southeast Asia, however, this type is poorly represented: it occurs only sporadically in Hmong-Mien and Tibeto-Burman (§3.5), and is entirely absent from Kra-Dai, Austroasiatic, and Austronesian (Fukushima et al. 2023). 2.6. Types with relative sex These types encode distinctions based on relative sex, that is, the sex of the speaker in relation to that of the referent. They are often combined with distinctions of relative age and/or the referent’s sex. This type can be subdivided into five subtypes: the Relative Sex Type (Type F), the Relative Sex/Age Type (Type FB), the Relative Sex/Skewed Age Type (Type FC), the Relative Sex/Age/Sex Type (Type FD), and the Relative Sex/Sex Type (Type FE). The Relative Sex Type is distinguished solely by relative sex. For instance, in Kiribati, a language of the Gilbert Islands in Melanesia, tari refers to a ‘same-sex sibling’ and m’āne to an ‘opposite-sex sibling.’ In this system, the same term has opposite values depending on the speaker’s sex: for male speakers, m’āne denotes ‘sister,’ whereas for female speakers it refers to ‘brother’ (Matsumoto 2000: 15). (6) Relative Sex Type in Kiribati (Marshall ed. 1983: 155) Same-sex Cross-sex tari m’āne Relative sex, when combined with distinctions of relative age and/or the referent’s sex, gives rise to more complex systems. One example is modern standard Korean, which exemplifies the Relative Sex/Skewed Age Type. In this system, terms for elder siblings vary depending on both the referent’s sex and the speaker’s sex, whereas younger siblings are referred to with a single undifferentiated term. (7) Relative Sex/Skewed Age Type in Korean (adapted from Fukui 2023: 19) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder hyɔŋ nuna oppa ɔnni Younger toŋsɛŋ 『地理言語学研究』 | Studies in Geolinguistics 5 44 Another example is found in the Hokkaido Ainu dialects, which follow the Relative Sex/Age/Sex Type. In this system, sibling terms encode distinctions of relative age, relative sex, and referent sex. While the terms for ‘elder brother,’ ‘elder sister,’ and ‘younger brother’ reflect the Age/Sex Type, the term for ‘younger sister’ additionally varies with the speaker’s sex. This pattern can be schematically represented as follows, based on our interpretation of Fukazawa (2023). (8) Relative Sex/Age/Sex Type in Hokkaido Ainu (adapted, ibid. p.11) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder yúp(-í) sá(-ha) yúp(-í) sá(-ha) Younger ák(-í) turés(-i) ák(-í) mátak(-i) Burushaski illustrates the Relative Sex/Sex Type, with three sibling terms that distinguish both the speaker’s sex and the referent’s sex, without encoding relative age (Yoshioka 2023). As shown in (9a), the terms vary depending on whether the speaker and sibling are of the same or opposite sex, but no distinction is made between elder and younger siblings. (9a) Relative Sex/Sex Type in Burushaski (adapted, ibid. p.44) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder -cọ -yás(t) -úlus -cọ Younger Since the system does not encode relative age, it can alternatively be represented in a simplified form, as shown in (9b). (9b) Simplified representation of the Relative Sex/Sex Type in Burushaski Same-sex Cross-sex Male referent -cọ -úlus Female referent -yás(t) Types involving relative sex occur with notable frequency in several language families and groups outside Southeast Asia, including Chukotko-Kamchatkan, Korean, Ainu, Ryukyuan, and Niger-Congo. They are also sporadically found in Japonic, Mongolic, Turkic, Dravidian, Nilo-Saharan, and various South Asian languages (Fukushima 2023, 2024). KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 45 Within Southeast Asia, although relative sex distinctions are absent in Sinitic sibling systems, they are found in varying degrees in Hmong-Mien, Kra-Dai, Austroasiatic, Tibeto-Burman, and Austronesian. In Hmong-Mien, this type is attested particularly in the western regions, contrasting with the dominance of the Age/Sex Type in the east (Taguchi & Tang 2023). In Austronesian, systems encoding relative sex are primarily found in the Pacific, as well as in Papua, the Solomon Islands, and the southern Philippines. These include the Relative Sex, Relative Sex/Age, Relative Sex/Skewed Age, and Relative Sex/Sex Types (Utsumi 2023). Languages in Kra-Dai, Austroasiatic, and Tibeto-Burman also show relative sexbased systems, though less frequently. A few Kra-Dai languages display the Relative Sex/Age/Sex Type. For instance, Laji, a Kra language spoken on the China-Vietnam border, uses different terms for ‘younger brother’ depending on the speaker’s gender. Similarly, Cun, a Hlaic language spoken on Hainan Island, marks relative sex in both ‘younger brother’ and ‘younger sister’ (Hirano et al. 2023). While Relative Age and Skewed Age Types dominate in Austroasiatic, several languages display relative sex distinctions. These include the Relative Sex Type in some Nicobaric languages; the Relative Sex/Age Type in Monic, Aslian, and Nicobaric; the Relative Sex/Skewed Age Type in Palaungic and Katuic; and the Relative Sex/Age/Sex Type in some Munda languages. Such systems are mostly found in coastal India, the Nicobar Islands, and the southern Malay Peninsula (Shimizu & Minegishi 2023). Tibeto-Burman also shows examples of relative sex-based types, including the Relative Sex/Age, Relative Sex/Age/Sex, and Relative Sex/Sex Types (see §§3.6–3.8). 3. Typology of sibling term systems in Tibeto-Burman This section presents a typological overview of sibling term systems in Tibeto-Burman, accompanied by illustrative examples. While some data come from primary sources collected through the authors’ fieldwork, many are drawn from secondary materials such as dictionaries and wordlists. Sibling term systems in Tibeto-Burman exhibit considerable typological diversity. The following types are identified in our data and will be discussed in turn: Undifferentiated Sibling Type (§3.1), Relative Age Type (§3.2), Skewed Age Type (§3.3), Age/Sex Type (§3.4), Sex Type (§3.5), Relative Sex/Age Type (§3.6), Relative Sex/Age/Sex Type (§3.7), and Relative Sex/Sex Type (§3.8). 3.1. Undifferentiated Sibling Type This type lacks lexical distinctions based on either relative age or the referent’s sex. Although it is rare in our Tibeto-Burman data, this is not entirely absent. Clear instances 『地理言語学研究』 | Studies in Geolinguistics 5 46 are found in Rawang and Wadamkong, two closely related Nungish languages. In Rawang, for example, the term nvm is used uniformly for siblings, referring to both male and female referents regardless of their relative age. (10) Rawang (LaPolla & Sangdong 2015: 287) Male Female Elder nvm Younger Relative age and referent’s sex can be specified compositionally by adding ageand sex-related morphemes to the neutral sibling term nvm. These include vlat ‘first,’ sv m ‘small,’ pè ‘male,’ and mè ‘female.’ All of these morphemes follow the neutral sibling term, as illustrated in (11). (11) Rawang (LaPolla & Sangdong 2015: 287) 3.2. Relative Age Type This type makes a single distinction based only on relative age. As noted by Matsumoto (2000: 7, 29), it is particularly concentrated in the Karenic branch of Tibeto-Burman, spoken in and around southeastern Burma. In our dataset, 26 out of 45 languages and dialects exhibiting this type are Karenic varieties. These include Blimaw, Bwe, Eastern Kayah Li, Geba, Pa-O, Pwo Karen, and Yingtalay, as well as various Kayan varieties, such as Diklon, Dokhoncon, Dolan, Dosanbu, Kadaw, Kalondei, Kathan, Lagu, Nangki, Nantwei, Phulon, Ramaku, Sonkan, Sonplao, Thaidai, Thaoku, and Zayein. An illustrative case is Eastern Pwo Karen, spoken in Karen State, Mon State, and the Tanintharyi region of Burma, in which sibling terms distinguish between elder and younger siblings without encoding the referent’s sex. (12) Eastern Pwo Karen (Kato 2004: 578) Male Female Elder wɛ  Younger phɯ  Male Female Elder nvm vlatpè nvm vlatmè Younger nvm svmpè nvm svmmè KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 53 (34) Tangam (Post 2017: 50) Male Female Elder aate aapoŋ Younger baʔpo bemme 3.5. Sex Type This type is characterized by a single distinction based on the referent’s sex. It is found in the Tibetic, Qiangic, Sal, and Digarish. These include Qiangic languages such as Hongding nDrapa, Longxi Southern Rma, Taoping Southern Rma, and Puxi Southern Rma; Usui, a Bodo-Garo language from the Sal group; Idu, a Digarish language spoken in and around Arunachal Pradesh; and a range of Tibetic varieties. One example comes from Usui, a Bodo-Garo language spoken in the Chittagong Hill Tracts of southeastern Bangladesh. (35) Usui (Huziwara 2010: 93) Male Female Elder tàkhu bәhãnàŋ Younger The Sex Type is relatively uncommon, attested in around 20 languages and dialects in our data, over half of which are Tibetic varieties such as Khyungkyog Tibetan, sTaglo Tibetan, and Rwatags Tibetan. (36) Khyungkyog Tibetan (Suzuki fieldnotes) Male Female Elder ɕa ȵe ʂɛ ɦu Younger 3.6. Relative Sex/Age Type The Relative Sex/Age Type encodes both relative age and relative sex. While this type is uncommon, our data include one instance. This is Xinyingpan Prinmi, a Qiangic language spoken in northwestern Yunnan, which employs distinct terms for same-sex siblings, while cross-sex siblings are referred to with a single term. 『地理言語学研究』 | Studies in Geolinguistics 5 54 (37a) Xinyingpan Prinmi (adapted from Ding 2014: 337, 348) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder pɜjR mʉLn ɥɛH mʉ Ln ɥɛH pɜjR Younger kɥɛF kɥɛF Since the Prinmi system employs only three terms, with some forms shared across categories, it can alternatively be represented more simply in terms of relative age and relative sex as follows. (37b) Simplified representation of the Prinmi system Relative sex Same sex Cross-sex Relative age Elder pɜjR mʉLn ɥɛH Younger kɥɛF 3.7. Relative Sex/Age/Sex Type This type involves distinctions based on relative age, relative sex, and the referent’s sex. It is attested in 23 languages and dialects in our data, including Lolo-Burmese languages such as Burmese and Nuosu, Sal languages such as Cak, certain languages of Arunachal Pradesh such as Kaman, and various Tibetic varieties. The Relative Sex/Age/Sex Type can be subdivided based on the number of lexical contrasts. For example, Dzongkha, a Tibetic language spoken in Bhutan, especially in the western valleys and central highlands, shows five lexical splits, as illustrated by the following example. (38a) Dzongkha (adapted from Imaeda 2006: 78) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder phogem 'azhim phogem 'azhim Younger nocu sim nocu num This system can alternatively be represented more explicitly in terms of relative sex, as shown in (38b) below. However, henceforth we will use speaker’s sex instead, since it provides a more intuitive representation. KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 55 (38b) Alternative representation of the Dzongkha system Relative sex Same-sex Cross-sex Referent’s sex M F M F Relative age Elder phogem 'azhim phogem 'azhim Younger nocu num nocu sim Myeik, a dialect of Burmese spoken in the southern coastal region of Burma, also exhibits five lexical splits, albeit with a slightly different configuration. In Dzongkha, it is the category ‘younger sister’ that is lexically distinguished, whereas in Myeik Burmese, the distinction falls on ‘younger brother’ instead, as shown in (39). Note that some male speakers use mau instead of ɲì, which suggests a shift toward the Age/Sex Type. (39) Myeik Burmese (adapted from Kato & Khin Pale 2012: 153) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder ʔakò ʔamǎ ʔakò ʔamǎ Younger ɲì ɲìmǎ ma  u ɲìmǎ Standard Burmese shows six splits, with younger siblings being referred to with four distinct terms depending on both the sex of the speaker and the referent. Elder siblings, by contrast, are referred to with the same two terms across speakers, as illustrated in (40). Note that in modern Burmese, hnămâ is frequently replaced by ɲìmâ, suggesting an ongoing shift toward a five-split system. (40) Standard Burmese (Kurabe fieldnotes) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder ʔăkò ʔămâ ʔăkò ʔămâ Younger ɲì hnămâ màuɴ ɲìmâ A similar six-term system is found in rGyaye Tibetan, an Amdo Tibetan variety spoken in Qinghai Province, China. (41) rGyaye Tibetan (Ebihara 2010: 84) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder awo atɕʰe awo atɕʰe Younger nu ʂaŋmo nyoŋwo nәmo 『地理言語学研究』 | Studies in Geolinguistics 5 56 Unlike rGyaye Tibetan, Themchen Tibetan, another Amdo Tibetan spoken in Qinghai Province, exhibits a further subdivision within the male elder sibling category based on relative sex. (42) Themchen Tibetan (Tournadre & Suzuki 2023: 728) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder hurgän ač’e apa ač’e Younger nu ȿangmo nyangwo nәmo The Relative Sex/Age/Sex Type can theoretically yield up to eight distinct forms, and at least one such fully elaborated case has been reported by Matsumoto (2000: 23– 24). Tournadre and Suzuki (2023: 726) add a new example, noting that one Amdo dialect has all eight terms. (43) Tsigorthang Tibetan (Tournadre & Suzuki 2023: 728–279) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder hurgän ȿangganma m nyangrgän ač’e Younger nu ȿangmo m nyangwo nәmo 3.8. Relative Sex/Sex Type In this type, lexical choices reflect relative sex and the referent’s sex, but are insensitive to age differences. This type is attested in five Qiangic languages in our dataset: Mätro nDrapa, Youlaxi (gYang-la-gshis) Choyu, Japhug, Wobzi Khroskyabs, and Siyuewu Khroskyabs. Tangut is also suggested to have exhibited this type (Jacques 2012). One example comes from Mätro nDrapa, a Qiangic language spoken in the western Sichuan region of China. (44a) Mätro nDrapa (Shirai fieldnotes) Speaker’s sex Male speaker Female speaker Referent’s sex M F M F Relative age Elder vɛȵʌpha1 neivʌ 1 mu 2 ȵʌȵʌpha1 Younger Since the system is not sensitive to age, it can be represented more simply, as in (44b). KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 57 (44b) A simplified representation of Mätro nDrapa Same-sex Cross-sex Male referent vɛȵʌpha1 mu 2 Female referent ȵʌȵʌpha1 neivʌ 1 A similar pattern is also found in Youlaxi Choyu, another Qiangic language spoken in Sichuan, although the corresponding sibling terms are not necessarily cognate. (45) Youlaxi Choyu (adapted from Huang ed. 2022: 152–154) Same-sex Cross-sex Male referent wә55phe55 mu 55 Female referent sqhɯ55phe55 ɕȵye 55 This sibling system is also attested in Wobzi Khroskyabs, a rGyalrongic language spoken in northwestern Sichuan, which employs four distinct terms. (46) Wobzi Khroskyabs (Yunfan Lai, p.c., 2023) Same-sex Cross-sex Male referent rmɑ stәɣ snәm Female referent sqʰæɕí mô In some languages, two distinct types coexist. For example, Japhug, another rGyalrongic language of northwestern Sichuan, exhibits both the Relative Sex/Sex Type, which encodes distinctions of relative sex and referent’s sex (47a), and the Relative Age Type, which marks only the age distinction between elder and younger siblings (47b). (47a) Relative Sex/Sex Type in Japhug (Jacques 2012: 216) Same-sex Cross-sex Male referent tɤ-xtɤɣ tɤ-snom Female referent tɤ-sqhɤj tɤ-wɤmɯ (47b) Relative Age Type in Japhug (Jacques 2012: 216) Male Female Elder tɤ-pi Younger ta-ʁi If two types coexist within a single language, we consider the language to exhibit both in Section 4. 『地理言語学研究』 | Studies in Geolinguistics 5 58 4. Geographical distribution and interpretation Tibeto-Burman languages, as illustrated in the preceding section, exhibit a wide range of typological patterns in their sibling term systems. Contrary to Murdock’s (1968: 7) suggestion that “the several types found in Tibeto-Burman are probably correlated with subfamilies,” this variation applies not only across different subgroups but also within a single subgroup. For example, as shown in Section 3, the Tibetic branch alone displays considerable diversity, including the Skewed Age Type (e.g., Lhasa Tibetan), the Age/Sex Type (e.g., Kyirong Tibetan), the Sex Type (e.g., Khyungkyog Tibetan), and the Relative Sex/Age/Sex Type (e.g., rGyaye Tibetan). Furthermore, the individual sibling terms are not necessarily cognate with one another.1 Given this diversity, it remains challenging to establish a single, definitive account of their historical development. That said, this section examines which patterns are more likely to reflect an older system within Tibeto-Burman, based on their genealogical and areal distribution. Figure 1 presents the geographical distribution of different types of sibling term systems in Tibeto-Burman. Among the attested patterns, the Undifferentiated Sibling Type, Relative Sex/Age Type, and Relative Sex/Sex Type are uncommon in our data, both in terms of geographical distribution and genealogical diversity. In other words, the first two types are attested in only a few languages, and the last is similarly restricted to a small number of Qiangic languages. Given their marginal status, these types are unlikely to reflect an older layer of the sibling term system within the family.2 In contrast to the three types discussed above, the Relative Age Type, Skewed Age Type, Age/Sex Type, Sex Type, and Relative Sex/Age/Sex Type are attested more broadly across the Tibeto-Burman area. These patterns appear across multiple branches and regions, suggesting greater stability and the possibility that they reflect an earlier stratum in the history of the family. The following section examines the historical status of each of these types within Tibeto-Burman sibling term systems. 1 For a detailed account of sibling term systems in Tibetic, see Tournadre and Suzuki (2023: 724–731). 2 Matsumoto (2000: 5) points out that the Undifferentiated Sibling Type is limited in its distribution across the world’s languages. Even in language groups where it is most frequently attested, such as the Philippine languages within the Austronesian family, it tends to emerge as an intermediate stage in the shift from one type to another, rather than reflecting an original system. K URABE , K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 59 Fig. 1 The geographical distribution of Tibeto-Burman sibling term systems Undifferentiated Sibling Type Sex Type Relative Age Type Relative Sex/Age Type Skewed Age Type (Elder-Skewed) Relative Sex/Age/Sex Type Skewed Age Type (Younger-Skewed) Relative Sex/Sex Type Age/Sex Type Within Asia, the Relative Age Type is particularly attested in Taiwan and in both Maritime and Mainland Southeast Asia (Matsumoto 2000: 6–7; Fukushima 2024). In our Tibeto-Burman data, it appears in 45 languages and dialects (§3.2). However, as noted by Matsumoto (2000: 7, 29), its distribution within Tibeto-Burman is notably concentrated in the Karenic branch, and in our dataset as well, more than half of the instances are from Karenic languages spoken in southeastern Burma. These languages are in close contact with Tai and Austroasiatic languages, where the Relative Age Type is widespread (Hirano et al. 2023; Shimizu & Minegishi 2023). This raises the possibility that the Karenic pattern represents a contact-induced innovation rather than an inherited feature. Comparable systems are also sporadically found in other branches such as Qiangic, Loloish, and Kuki-Chin-Naga. However, they are not dominant in these groups, where alternative sibling term systems are attested. For example, Qiangic languages show not 『地理言語学研究』 | Studies in Geolinguistics 5 60 only the Relative Age Type but also Skewed Age, Age/Sex, Relative Sex/Age, and Relative Sex/Sex Types (§3). Likewise, some languages such as Japhug exhibit both the Age-based and the Relative Sex/Sex Types, which may suggest a transition from one type to another (§3.8). Outside the Karenic branch, thus, the pattern may have arisen similarly through language contact or through internal simplification of more complex systems. If this is the case, the Relative Age Type may not reflect an older stratum of sibling term systems within Tibeto-Burman. Fig. 2 The geographical distribution of Tibeto-Burman sibling term systems (Detailed) The Sex Type (or the Brother/Sister Type) is attested in approximately 20 languages and dialects in our dataset, which belong to the Tibetic, Qiangic, Sal, and Digarish groups. However, in some of these cases, particularly those found in the Indosphere, 3 the type may have developed secondarily through language contact with Indo-Aryan languages. This interpretation is supported by a typological trend in the world’s languages noted by Matsumoto (2000: 11–15), who observes that the Sex Type is strongly associated with the Indo-European and Afro-Asiatic families, and that instances of this type outside these families are often explained as the result of language contact with them. Matsumoto (2000) further observes that languages with this type typically have grammatical gender. This suggests that the type is unlikely to be 3 For the conceptual distinction between the “Sinosphere” and the “Indosphere” in the context of language contact and cultural influence in Southeast Asia, see Matisoff (1990). KURABE, K. et al.: A typology and geolinguistics of sibling term systems in Tibeto-Burman 61 typologically compatible with Tibeto-Burman, in which grammatical gender is generally absent. The Skewed Age Type is attested across a wide range of Tibeto-Burman branches and does not exhibit any strong areal bias. It is found in diverse groups, including LoloBurmese, Qiangic, Tibetic, Tujia, Sal, Kuki-Chin-Naga, Tani, Kiranti, and Newaric. Its distribution, both genealogically and geographically broad and balanced, suggests that it may represent an older stratum within the sibling term systems of Tibeto-Burman. Note that the Skewed Age Type is also widely attested in Austroasiatic languages, appearing across several branches including Bahnaric, Katuic, Khmuic, Mangic, Monic, Palaungic, and Vietic (Shimizu & Minegishi 2023). Given the long-standing presence of Austroasiatic in Mainland Southeast Asia (Bellwood 1992), some Tibeto-Burman instances of this type may have arisen through language contact. This type is also attested in Kra-Dai, where its distribution is confined to Northern Tai (especially northern Zhuang dialects), Central Tai, and the Li languages of Hainan Island (Hirano et al. 2023). Some Tibeto-Burman instances of this type may similarly have arisen through local contact with these languages. The Age/Sex Type is one of the most prevalent sibling term systems crosslinguistically and is notably stable in both internal structure and areal distribution (Matsumoto 2000: 10). This type, as described in Section 3.4, is quite common in Tibeto-Burman, both genealogically and geographically, and is attested in a wide range of groups. This widespread distribution suggests that it may represent one of the oldest sibling term systems in Tibeto-Burman. However, some instances may have arisen secondarily through language contact. Notably, this type is found throughout Sinitic, with a long historical continuity (Matsumoto 2000: 10). Fukushima (2024: 153) suggests that this type may have spread to other language families through contact with Sinitic. In this context, its presence in Tibeto-Burman languages, particularly in those within the Sinosphere, should be considered in relation to possible Sinitic influence. Additionally, Matsumoto (2000: 11) notes that the Age/Sex Type is closely related to the Skewed Type, and that in many languages there is a systematic shift from the latter to the former, with the two sometimes coexisting without a clear distinction. These observations highlight the need to consider the possibility of internal development as well. The Relative Sex/Age/Sex Type is found in Lolo-Burmese, Sal, Tibetic, and some languages of Arunachal Pradesh. Matsumoto (2000: 29–30) suggests that this type reflects an earlier stage in Tibeto-Burman, based on its sporadic distribution across the region and on Benedict’s (1942: 314) description of Classical Tibetan, which he regards as exemplifying this type. He further argues that sibling terminology in Tibeto-Burman has shifted from an earlier pattern structured around relative sex distinctions toward the 『地理言語学研究』 | Studies in Geolinguistics 5 62 Age/Sex Type, where both age and sex are involved, with the role of relative sex gradually diminishing in importance. The Myeik Burmese data discussed in Section 3.7 could be situated as an example of this shift. Figure 1 also suggests that this type is scattered around the periphery of the Tibeto-Burman area, consistent with Matsumoto’s (2000) observation. However, identifying types involving relative sex is not always straightforward when relying on secondary sources, and additional empirical data will be required for more conclusive analysis. 5. Conclusions This study has examined the typology and geolinguistic distribution of sibling term systems in Tibeto-Burman. The systems observed in our data display considerable diversity, encompassing the following types: Undifferentiated Sibling Type (§3.1), Relative Age Type (§3.2), Skewed Age Type (§3.3), Age/Sex Type (§3.4), Sex Type (§3.5), Relative Sex/Age Type (§3.6), Relative Sex/Age/Sex Type (§3.7), and Relative Sex/Sex Type (§3.8). Our data further demonstrate that this variation occurs not only across different subgroups but also within individual subgroups. Such typological variation reflects broader patterns of lexico-semantic diversity, paralleling those found in the lexicon and grammar of Tibeto-Burman languages (e.g., Matisoff 1978, 2003; Thurgood & LaPolla 2017). While it is difficult to draw a unified historical conclusion from this diversity, certain types, such as the Undifferentiated Sibling Type, the Relative Sex/Age Type, and the Relative Sex/Sex Type, can be regarded as secondary, given their skewed genealogical and geographical distributions. Other types, particularly the Skewed Age Type, the Age/Sex Type, and the Relative Sex/Age/Sex Type, may reflect older strata, as suggested by their genealogical diversity and geographically non-contiguous distribution. Nevertheless, given the pervasive and often intensive nature of language contact across the Tibeto-Burman area, as well as the potential for independent internal developments, both external and internal factors should always be taken into account. This issue extends beyond the scope of a broad typological and geolinguistic survey such as the present study, and further progress will require detailed descriptive studies grounded in solid fieldwork. Furthermore, much of the data used in this study derives from secondary sources, and the reliability of interpretation is therefore a constant concern. This is particularly true for complex systems, such as those involving relative sex distinctions, which are not always accessible through secondary materials. Even when the original sources were based on primary data, the influence of contact languages may have obscured the