Chapter 6.2 Kinship

Malcolm Ross and Jeffrey C. Marck

1. Introduction¹⇫

Kinship terms are terms used to address or refer to one’s relatives, e.g. ‘mother’, ‘son’, ‘grandfather’ and the like. Unlike the lexical items discussed elsewhere in these volumes, a language’s kinship terms form a more or less closed system. The system determines what distinctions its terms encode, and the definitions of terms are largely constrained by the system they belong to—‘largely’ because a system’s terms sometimes overlap in meaning. The goal of this chapter is to reconstruct early Oceanic kinship terms and to hypothesise the system to which they belonged.

A number of linguists and anthropologists have treated kinship terms as a window into the social organisation of present and past kinship systems. Works that reconstruct Oceanic social organisation in this way include Rivers (1914), Blust (1981c), Pawley & Green (1984), Hage (1999, 2001a, ch.3 of this volume) and Turner (2007). Moving further back in time, early Austronesian kinship systems are reconstructed by Blust (1980a, 1980c, 1994).

In this chapter we touch on the reconstruction of social organisation only as it relates to lexical reconstruction. A kinship terminology and the social organisation of its users are often not in a one-to-one relationship. A terminology may retain terms that reflect past social conditions. For example, a number of Oceanic kinship terminologies indicate that their communities once practised a form of cross-cousin marriage whereby a man’s preferred potential wife was his mother’s brother’s daughter. (§2.4.1.2.5). Some communities still practised it at first contact with Europeans. Others apparently did not, yet the terms implying it still persisted, perhaps because the terminology continues to enshrine what members of the community regard as an ideal (McKinley 1971). A paper on the matrilineality of early Oceanic social structure by the late Per Hage, an anthropologist with linguistic interests, is reprinted as chapter 3 of this volume.

This chapter builds on earlier attempts to reconstruct a POc kinship terminology. Wilhelm Milke’s 1938 paper on POc sibling terms built on a collection of data remarkable at the time, and was expanded twenty years later (Milke 1958b) into a full reconstruction of the POc terminology. He further added to it in later papers (Milke 1965:345–346; 1968:158–161, 167). Pawley (1981) and Pawley & Green (1984) provide a listing of POc kinship terms which largely follows Milke’s. Chowning (1991) includes a critique of previous reconstructions and presents a modified version of the POc terminology. The present chapter is based on a fresh analysis of a larger database that digitisation makes possible, with the consequence that its results sometimes differ from its predecessors.²

Qualitatively, the data collected for this chapter fall into three rough categories. In the first category are the reasonably comprehensive accounts of kinship terminologies and associated practices that are found in some ethnographies. The second category comprises incomplete terminologies constructed from lists of kinship terms in some ethnographic writings and a few ethnographically informed dictionaries and grammars. In the third category are individual terms collected from dictionaries and wordlists. These are often glossed only vaguely, e.g., ‘uncle’, ambiguous because Oceanic languages tend to label a paternal uncle (father’s brother) and a maternal uncle (mother’s brother) differently. The sources of the first two data categories are listed in an appendix at the end of the chapter. The sources of the third kind of data are those used in chapters throughout these volumes, listed in Appendix A towards the back of the volume.

This chapter falls into four parts. The introduction continues in §2.2 with a description of two Oceanic terminologies, Nakanai and NE Ambae, providing a framework for what follows. The second (§2.3) is a typological overview of Oceanic kinship terminologies, based as far as possible on comprehensive terminologies. This leads to §2.3.5, describing the probable structure of the POc kinship terminology and the forms of its terms. The last and longest part (§2.4) reconstructs POc kinship terms themselves.

2. Introducing Oceanic kinship terminologies⇫

This section examines the terminology in Nakanai,³ a MM language of New Britain, as an example of the distinctions often encoded in Oceanic terminologies. Figure 2.1 shows the Nakanai terms for blood relatives, Figure 2.2 for in-laws.

Before turning to Nakanai, two matters need an explanation. First, as explained in vol.1:32 and in vol.5:75-76, most Oceanic languages outside Polynesian⁴ reflect a POc grammatical distinction between directly and indirectly possessed nouns (Lichtenberk 1985a). A directly possessed POc noun took a possessor suffix, marked here with a following hyphen (e.g. *tama- ‘father’: *tama-gu ‘my father’, *tama-mu ‘your (singular) father’, *tama-ña ‘his/her father’ etc), whereas an indirectly possessed noun required no suffix. Directly possessed nouns are inalienable; that is, they are items that usually do not exist without a possessor. They include many kinship terms. But the converse is not true: not all kinship terms are directly possessed.

Secondly, in the middle of the figures is a person labelled ‘EGO’. EGO (Latin ‘I, me’) is ‘the speaker’, i.e. the person in relation to whom some other person, ALTER (Latin ‘other’), is labelled. Thus if I am EGO, my father is an ALTER labelled in relation to me. If he is defined in relation to some other EGO, he might then be my brother, my son, my grandson, and so on.

2.1. Attributes encoded in a kinship terminology⇫

Figure 2.1 shows that a male EGO distinguishes between brothers who are older than him (tua-) and those who are younger (tari-), whereas English has no term with a meaning that depends on the age of ALTER relative to EGO.

Figures 2.1 and 2.2 show Nakanai terms from the perspectives of both a male and a female EGO.⁵ The differences between them take two forms. First, a male EGO refers to his brother as tua- or tari-, whereas a female EGO uses tua- or tari- for her sister. That is, Nakanai tua- and tari- must be glossed ‘sibling of same sex as EGO’ (usually abbreviated ‘same-sex sibling’ or ‘s.s. sibling’), not ‘brother’ or ‘sister’. The sex of tua- or tari- depends on EGO’s sex. Second, a male EGO labels his opposite-sex sibling hatavile-, but a female EGO refers to her opposite-sex sibling as hatamale-. We prefix the conventional ‘female’ and ‘male’ symbols to a gloss to indicate who uses it: thus hatavile- ‘♂sister’ and hatamale- ‘♀brother’. In English, however, no distinction between terms depends on EGO’s sex. The distinction between English wife and husband encodes ALTER’s sex, not EGO’s. Indeed, ALTER’s sex plays a significant role in the English terminology, with six basic pairs: mother/father, aunt/uncle, sister/brother, daughter/son, niece/nephew and wife/husband. Nakanai has only one: tila-/tama ‘mother/father’. Note, too, that EGO distinguishes among s.s. siblings by age relative to EGO, but labels o.s. siblings by their (ALTER’s) own sex.

In the figures and elsewhere, generations are indicated with, e.g., G₀ ‘EGO’s generation’, G₊₁ ‘one generation above EGO’, G₋₂ ‘two generations below EGO’, and so on. In Figure 2.1, tubu- appears against G₊₂, indicating that all members of G₊₂ are tubu- in Nakanai. It also exemplifies the classificatory nature of Oceanic kinship terms. The term tubu- labels all members of G₊₂. Most kinship terminologies include classifying terms. English has some too: uncle labels EGO’s father’s brothers, father’s sisters’ husbands, mother’s brothers and mother’s sisters’ husbands. Aunt and cousin are also classifying. Nakanai and English classify siblings-in-law in the same way: English brother-in-law refers to EGO’s wife’s brother and to EGO’s sister’s husband, and Nakanai iva- does the same.

Some kinship terms in Oceanic languages are used reciprocally. A Nakanai EGO refers to his/her mother’s brother as hala- and, when EGO is a child’s hala-, he refers to his sister’s child as hala-. That is, the term is reciprocal across generations. Figure 2.1 also shows that some grandparents and grandchildren refer to each other reciprocally as tubu-.

The Nakanai terminology classifies blood relatives of G₊₁ in a particular way. EGO’s father’s brother is labelled with the same term as EGO’s father, and EGO’s mother’s sister with the same term as EGO’s mother. This has two consequences. One is that father’s brother’s descendants have the same labels as father’s descendants, and mother’s sister’s descendants the same labels as mother’s. The other consequence is that what in English would be cousins fall into two Nakanai categories. Father’s brother’s and mother’s sister’s children are labelled as if they were EGO’s siblings, but father’s sister’s and mother’s brother’s children have a different label, lavo-. In the kinship literature the former are labelled ‘parallel cousins’, the latter ‘cross-cousins’. ‘Cross’ here refers to the fact that cross-cousins are related to EGO via opposite-sex parents. The classifying attribute of Oceanic kinship terms is discussed further in §2.2.3.

The last few paragraphs have touched on various attributes of kinship terminologies in general. Their presence or absence in Nakanai and English is summed up in Table 2.1. Presence is shown by a tick. A particular attribute may or may not be reflected in a terminology.⁶ For example, English has no terms that depend on sex relative to EGO.

‘Sex of linking relative’ refers to the contrast between, e.g. ‘father’s brother’ and ‘mother’s brother’.

Table 2.1 Attributes and values of Nakanai and English terms

Attribute	Nakanai	English	values of terms
generations distinguished	✔	✔	–2, –1, 0, +1, +2 etc
sex of ALTER	✔	✔	male, female, either
sex relative to EGO	✔		same, opposite, either
sex of linking relative	✔		same, male, female
seniority within generation	✔		elder, younger, none
reciprocity across generations	✔		yes, no
affinity	✔	✔	blood relative, in-law

2.2. ‘Parallel’, ‘cross’ and kinship notation⇫

As noted above, ethnographers use the terms ‘parallel cousin’ for cousins related to EGO via mother’s sister or father’s brother and ‘cross-cousin’ for those related to EGO via mother’s brother or father’s sister.⁷ Similar terms are sometimes used of siblings, such that a ‘s.s. sibling’ is a sibling of the same sex as EGO and a ‘cross-sibling’ or ‘cross-sex sibling’ is of the opposite sex to EGO. This usage is inconsistent, as ‘parallel cousin’ and ‘cross-cousin’ refer to their parent’s sex, not their own, and the latter sometimes needs to be stipulated as in ‘same-sex cross-cousin’ or ‘opposite-sex cross-cousin’. For consistency’s sake, we follow Murdock 1949, 1968a). replacing ‘cross-sibling’ by ‘opposite-sex sibling’ (‘o.s. sibling’).

Table 2.2 Abbreviating kinship terms

M	mother	(NB not ‘Male’)
F	father	(NB not ‘Female’)
P	parent

Z	sister	(Z, as S is reserved for ‘son’)
B	brother
G	sibling	(Latin germāna, germānus)

W	wife
H	husband	(French épouse, époux)
E	spouse

D	daughter
S	son
C	child

In the interests of space, glosses in cognate sets (§2.4) and figures employ a version of the conventional ethnographic abbreviations for kinship terms. Its basic building blocks are shown in Table 2.2. In each of the four sub-table, the third abbreviation, an innovation borrowed from Hage (2001b), embraces the two categories above it. These capture Oceanic categories and reduce abbreviations so that, for example, ‘parent’s sibling’ is expressed as ‘PG’, instead of the more conventional ‘FB, MB, FZ, MZ’. The syntax of these terms is straightforward, e.g. MZ ‘mother’s sister’, EGC ‘spouse’s sibling’s child’, and so on.

There are four modifiers, each abbreviated as a lower-case letter preceding one of the letters in Table 2.2. They are:

s	same sex as EGO
o	opposite sex to EGO
y	younger than EGO within EGO’s generation
e	elder than EGO within EGO’s generation

Thus oC ‘opposite sex child’; ysG ‘younger same-sex sibling’.

Relative sex is reckoned relative to EGO in order to avoid ambiguity. Hence EGsC ‘spouse’s sibling’s child of EGO’s sex’, not ‘spouse’s siblings’s child of spouse’s sibling’s sex’. However, there are terms that encode sex relative to someone other than EGO, and in these cases curly brackets are used, e.g. {PsG}C ‘child of parent’s same-sex sibling’,⁸ i.e. ‘parallel cousin’, as opposed to PsGC ‘child of parent’s sibling of EGO’s sex’.

2.3. Nakanai terms and their definition⇫

Like a good many other societies in Oceania, Nakanai society has divisions based on descent (Chowning 1965 calls them ‘sibs’). Some societies, like Tolai (Fingleton 1986), have just two divisions or ‘moieties’. Others, like Nakanai, have several divisions. In many Oceanic societies descent is matrilineal (see ch 3). In other words, a child belongs to its mother’s lineage. Since marriages are between members of different divisions, this means that a father’s children belong to his wife’s division. Among EGO’s closer relatives the only members of EGO’s generation who are of another division are cross-cousins. Being of a different division from EGO, they are possible marriage candidates.

Nakanai kin terms for members of EGO’s generation are shown in Table 2.4. The values of the four relevant attributes (see Table 2.1) are shown in the four rightmost columns. Each term is defined by a unique set of values. The cross-cousin term lavo- encodes neither referent’s sex nor relative sex. Its salient feature is a social one: a lavo- can be married by EGO.

The terms for members of G₊₁ and G₋₁ are shown in Table 2.4. Of the seven terms, four are reciprocal, giving two entries, ‘+1’ and ‘–1’ in the ‘Generation’ column.’ Notably the terms in Table 2.3 are defined by a different set of attributes from those in Table 2.4, i.e. they make different distinctions.

Terms at G₊₁ in Figure 2.3 and at G₋₁ in Figure 2.4 follow a rule. The wife of any tama- is a tila- and vice versa. The wife of any hala- is an isa- and vice versa. This accounts, for example, for the fact that in Figure 2.3 father’s sister’s husband is labelled hala- and mother’s sister’s husband tama-.

Table 2.3 Nakanai kinship terms of EGO’s generation

Table 2.4 Nakanai kinship terms of EGO’s generation

In Figure 2.1 there is just one reciprocal term for members of G₊₂ and G₋₂. This is tubu- ‘grandparent/grandchild’. However, at G₋₂ tubu- is restricted to the grandchildren of EGO’s s.s. siblings and s.s. parallel cousins, whilst the grandchildren of o.s. siblings and o.s. parallel cousins are labelled lavo-. This also seems to be rule-governed. Because Nakanai society is matrilineal and EGO’s spouse is of a different division from ego, male EGO’s latu- ‘children’ are not of his lineage or division, but his sister’s children, hala-, are. The situation is reversed if EGO is female: latu- are of her division, but isa- are not. The rule then stipulates one uses tubu- for the child of one’s latu-, but lavo- for the child of one’s hala- or isa-.

The application of tubu- to the children of one’s cross-cousins, i.e. relatives of the same generation as one’s children, not one’s grandchildren, is obscure, but is touched on in §2.3.3.

In many Oceanic cultures the mother’s oldest living brother has the greatest authority within his matriline to instruct his sister’s children on societal matters and to make decisions about matters such as their marriages. The way in which this authority relates to that of the father, who does not belong to his children’s division, varies from society to society. The role of the mother’s oldest brother means that there is typically a dedicated term for mother’s brothers. In Nakanai this is hala-. The fact that in many Oceanic cultures mother’s oldest brother had or has significant responsibilities towards his sister’s children (♂ZC) is enshrined in a special term that a male EGO uses to refer to those children. It also captures the fact that in a matrilineal society a man’s sister’s son is his heir. The special term for sister’s son/sister’s child is sometimes the term for MB, used reciprocally, as it is in Nakanai (see further §2.3.2.2).

Figures 2.1 and 2.2 and Tables 2.3 and 2.4 are unavoidably limited in scope, and the classificatory senses of the terms are not presented fully. Classifying terms are wide in scope. Indeed, most cover a theoretically unbounded set of relationships. How can one capture the sets that these glosses represent? For tama-, for example, an exhaustive definition might be:

• male blood relative of EGO’s father’s generation, related to EGO through EGO’s father’s male ancestors and their male descendants; and
• husband of female blood relative of EGO’s mother’s generation; the female relative is related to EGO through EGO’s mother’s female ancestors and their female descendants.

This definition includes ‘father’ and ‘father’s brother’ as primary referents. Both are grandfather’s sons, or FFS. They and grandfather’s brother’s son are great-grandfather’s grandsons—and so on. Tracking back through ego’s ancestry using the first part of the definition gives a set of glosses for tama-: FFS, FFFSS, FFFFSSS and so on, abbreviated as FnFSn, where the two instances of n are equal. Doing the same with the second part gives MnMDnH. Glosses of this kind, not used elsewhere in this chapter, serve to show that classifying terms denote theoretically unbounded but definable classes in many Oceanic languages, and are in this respect fundamentally different from English. The Tolai (MM), for example, who number in the tens of thousands, see themselves as all related⁹ and as members of one of the two Tolai divisions (moieties), and therefore as having a kin relationship with every other Tolai (Epstein 1969:122; Fingleton 1986:304).

Ethnographers argue over whether a term like Nakanai tama- has one meaning or more than one.¹⁰ Those who argue for two or more meanings (polysemy) distinguish between a narrow sense, ‘father’, and a classificatory sense, FₙFSₙ (Malinowski 1929:442, 495–96, 513, 527; Lounsbury 1965; Scheffler 2003; Shapiro 2018). Those who argue for one meaning (monosemy) regard something like the definitions above as the appropriate gloss of tama- (Leach 1958, attacking Malinowski; Schneider 1984). The question for a linguist is, What would a dictionary definition look like, and why? Pawley & Sayaba’s draft dictionary of Wayan Fijian has the following (edited and reordered):

tama- 1. Father. 2. Classificatory father, one who is brother or classificatory brother to one’s father. 3. Uncle, specifically the husband of one’s mother’s sister or classificatory sister.

This is clearly a polysemous definition. It is perhaps a linguistic universal that speakers often use a term with a narrow sense for a larger class of which the narrow-sense referent is perceived as the most salient member. How this occurs in folk classifications of plants and animals is discussed by Evans (2008) and Pawley (2011b) respectively. The definition above is an example of this broad phenomenon in the context of a kinship terminology.

Ethnographic accounts indicate that, for example, ‘father’ in the narrow sense typically has a relationship with his offspring that is closer to them than that of his brothers and other classificatory ‘fathers’, and that speakers certainly understand and act on a difference between the two senses (Codrington 1891:36–37; Humphreys 1926:34; Malinowski 1929:442–443; Powdermaker 1933:137–138; Mead 1934:220; Blackwood 1935:59–60; Fischer 1966:117; Hogbin 1964a:17; Scheffler 1965:75–76; Wagner 1986: xvi, 62).

This issue leads into another ethnographic question. Are definitions of kinship terms cognitively real? Do they reflect speakers’ concepts? Goodenough (1956, 1965) answered ‘yes’; Burling (1964) vigorously ‘no’. A dictionary-maker would argue that dictionary definitions should be a good approximation of cognitive reality, otherwise speakers cannot agree (or disagree) with them.

3. The typology of Oceanic kinship terminologies⇫

The typology of Oceanic kinship terminologies briefly presented here is based on 139 cases where a complete or near-complete terminology is available covering ego’s parents and their siblings and the parents’ and siblings’ children. Ideally each terminology should also include all their children, but many otherwise detailed ethnographies omit terms for the children of some or all of ego’s cousins. The terminologies also cover members of the grandparent and grandchild generations. Sources are in the appendix to this chapter.

These 139 languages are referred to below as ‘the sample’, but they are not a sample chosen for statistical purposes. They are simply the languages for which relevant data are available. NCV is under-represented, New Caledonia and Fiji over-represented. Nonetheless, they are distributed throughout Oceanic.

The typological investigation in §2.3.1–4 does not cover in-law terms.

3.1. Bifurcate merging and its modifications⇫

The Nakanai terminology in §2.2.3 exemplifies the fact that a terminology’s semantic structure determines the meaning of each term. Conversely, each term covers one or more kinship slots, e.g. tama- covers F, FB and MZH (and more). If the slots covered by each term laterally as far as and including ego’s parents’ siblings are diagrammed, we have a visual representation of a terminology’s structure, as in Figures 2.3 and 2.4.

Lowie (1928) categorised kinship terminologies, based on how they treat G₊₁. He called terminologies like Nakanai ‘bifurcate merging’—‘bifurcate’, because the terms for ‘mother’ and ‘father’ are distinct, and ‘merging’ because each parent and her/his s.s. siblings are merged terminologically (called by the same term), but o.s. siblings are labelled differently (§2.2.3).

Lowie’s ‘generational’ category, where all uncles and aunts are treated as parents, turns up in a sprinkling of Oceanic languages (see later in this section).

English, on the other hand, Lowie classed as ‘lineal’ because terms for ego’s ancestors and descendants (‘lineal kin’), as well as siblings, are distinct from those for uncles, aunts and cousins (‘collateral kin’), which are all categorised by generation. There are no Oceanic lineal terminologies.¹¹

Murdock (1949) extends Lowie’s definitions by including G₀ as well as G₊₁. Thus a bifurcate merging terminology is one which is

1. bifurcate: there are distinct terms for ‘mother’ and ‘father’ (M ≠ F);
2. merging:
   1. EGO’s father and father’s brother are both called ‘father’ and EGO’s mother and mother’s sister are both called ‘mother’ (FB = F, MZ = M);
   2. EGO’s father’s brother’s and mother’s sisters children, EGO’s parallel cousins, are called by the same terms as EGO’s siblings (FBC = MZC = G).
3. laterally restricted:
   1. EGO’s mother’s brother is called by a term other than ‘father’ and EGO’s mother’s sister by a term other than ‘mother’ (MB ≠ F, FZ ≠ M);
   2. EGO’s mother’s brother’s and father’s sisters children, ego’s cross-cousins, are called by terms other than those for ego’s siblings (MBC ≠ G, FZC ≠ G).

The grey rectangle in Figure 2.3 encloses the part of the terminology that is ‘merged’.

A working definition of ‘merging’ is needed here. Sources sometimes give binomials (terms consisting of two lexical items), like Sye (SV) drinme- pelaɣ ‘mother’s sister’, where drinme- is ‘mother’ and pelaɣ apparently means ‘removed’, ‘apart’ or, with kinship terms, ‘adoptive’ (there is no pelaɣ entry in Crowley 2000). Because the second, modifying, part of many binomials is omitted in daily use, a kinship binomial like drinme- pelaɣ is counted for analytic purposes as ‘the same’ as drinme- ‘mother’, i.e., as merged.

Of the 139 sample languages, 65 (47%) are bifurcate merging in accordance with the definition above.

At the core of a bifurcate merging structure are the mergers in (2) above and, remarkably, they are found in 137 of the 139 sample languages. One language, Sye, lacks one of the mergers in (2a), and two, Sye and Nadrau Fijian, lack (2b).¹² However, whereas the mergers in (2) are near-universal in Oceanic, there are numerous languages that do not conform to the definitions in (3), because at least one term for a merged category is extended to include a G₊₁ or G0 term outside the rectangle. In some languages the restriction in (3a) is breached, in others the restriction in (3b), and in yet others both restrictions in (3). The languages in which each breach occurs are shown in Table 2.5.

Table 2.5 Oceanic terminologies with bifurcate merging and breaches thereof

	Bifurcate merging	FZ = M	FZ = M and MB = F	PGC = G
Adm	Mussau, Sori, Lele, Pak, Baluan, Lou	Yapese	—	—
SJ and NNG	Yabem, Bukawa, Middle Watut	Sobei, Wogeo, Mapos Buang	—	o.s: Takia, Adzera, Mapos Buang
PT	Kilivila, Muyuw, Dobu, Galea, Sudest	Kalauna, Tubetube	N Mekeo	Bwaidoga, Sinaugoro (Saroa), Motu, Roro, N Mekeo
MM	Nakanai, W Kara, Lakurumau, Nalik, Lihir, Notsi, Lamusong, Lelet, Usen Barok, Sursurunga, Patpatar, Tolai (both)	Nehan, Petats, Hahon, Mono-Alu, Nduke, Roviana, Marovo, Varisi	Simbo	Vitu (o.s.), Mono-Alu, Varisi, Simbo, Nduke, Roviana
SES	NWGR,¹ Baegu, E Arosi	Bugotu, Gela, Lengo, Birao, Longgu, Kwaio, Sa’a, Fagani, W Arosi, Owa	—	W Guadalcanal (Gae), Lengo, To’aba’ita, Kwaio, Fagani, W Arosi, Owa
NCV	Araki, Raga, Big Nambas, N Ambrym, Nguna	Loh, Mota, Akei, NE Ambae, Nduindui, Vao, Ninde, Sinesip	—	Loh
SV	Kwamera	Anejom	—	—
NCal	Belep, Nyelâyu, Yuanga, Kumak, Fwai, Pwaamei, Bwatoo, Wahmwaang, Cèmuhî, Paicî, Arhâ, Ajië, Xârâcùù, Xaragure, Numèè, Kwenyii	—	—	Nengone, Drehu, Iaai
Mic	—	Nauruan, Marshallese, Satawalese	Chuukese	Marshallese (s.s.), Satawalese, Chuukese, Lamotrek
Fij	Wayan, Vuda, NW Viti Levu, Deuba, Kadavu, Bauan, Nadrau, Tokatoka, Moala, Matailobau, Lakeba	Cakaudrove, Macuata	—	Tavua, MacuataTokatoka
Pn exc EPn	Ifira-Mele, W Futunan	Rennellese, Takuu, Luangiua, Sikaiana	Pukapuka	Tongan, Tokelau, E Futunan, Rennellese (s.s.), Anuta, Tikopia, Pileni, Luangiua, Sikaiana
EPn	—	—	Tahitian, Māori, Rapa, Hawaiian	Rapa, Maori, Tahitian, Marquesan, Hawaiian

¹ NWGR = Rivers’ (1914) Northwest ‘Guadalcanar’.

Taking (3a) first, two breaches occur:

1. EGO’s father’s sister is called by the term for mother/mother’s sister (FZ = M: 44 cases);
2. EGO’s father’s sister and mother’s brother are called respectively by the terms for mother/mother’s sister and father/father’s brother (FZ = M, MB = F: 8 cases);

The case numbers tells us that FZ = M occurs far more often than MB = F. Furthermore, there is an implicational relationship such that MB = F only occurs where FZ = M occurs, but never vice versa. The distinctiveness of the ‘mother’s brother’ category is almost as durable as the mergers in (2), whereas the ‘father’s sister’ category is readily lost.

However, the loss of an FZ term has perhaps not occurred as frequently as Table 2.5 might imply. The MM languages split into two groupings. Those listed under ‘Bifurcate merging’ are all located in New Britain and New Ireland. Apart from Vitu, those listed in the other three columns are all members of the well established NW Solomonic subgroup (Ross 1988, chapter 7), and the innovation extending POc *tina- ‘mother’ to MZ probably happened in Proto NW Solomonic and was inherited by its member languages.

Turning to (3b), three breaches occur:

3. EGO’s cross-cousin is called by a term for sibling/parallel cousin (MBC = FBC = FZC = MZC = G, i.e. PGC = G); this may apply to
   • only s.s. cross-cousins (2 cases);
   • only o.s. cross-cousins (4 cases);
   • all cross-cousins (40 cases).

Terminologies in which the breach applies only to s.s. or only to o.s. cross-cousins are shown in italics in the rightmost column of Table 2.5.¹³

Each of these breaches brings a system nearer to Lowie’s generational category. The outcome of (i) above is that all female members of G₊₁ are called ‘mother’, of (ii) that all members of G₊₁ are called ‘mother’, ‘father’ or ‘parent’. The outcome of (iii), when it applies to all cross-cousins, is that all members of G₀ are called by sibling terms. Note that (i), (ii) and (iii) do not affect distinctions of sex (absolute or relative-to-ego) or relative age that a terminology may make.

Table 2.6 shows the terminologies with tendencies toward a generational structure. Column 4 adds (1), M = F, i.e. a uninomial term for ‘parent’, with ‘mother’/‘father’ distinction. A terminology that lacks this distinction is not bifurcate. There is no implicational relationship between the breaches in (2) and (3) above, but there is a strong tendency that if in a language FZ = M & MB = F, then PGC = G (Table 2.5). The converse is not true.

Of the 46 languages where PGC = G (Table 2.5, rightmost column), 22 have no other tendency toward a generational structure (Table 2.6, column 1).

Table 2.6 Oceanic kinship terminologies that tend toward Lowie’s generational structure

	1	2	3	4
	Only PGC = G	Only PGC = G and FZ = M	PGC = G and FZ = M and MB = F	PGC = G and FZ = M and MB F and M = F
Adm	—	—	—	—
SJ and NNG	Takia, Adzera (o.s.)	Mapos Buang (o.s.)	—	—
PT	Bwaidoga, Sinaugoro (Saroa), Motu, Roro	—	N Mekeo	—
MM	Vitu (o.s.)	Mono-Alu, Nduke, Roviana, Varisi	Simbo	—
SES	W Guadalcanal (Gae), To’aba’ita	Lengo, Kwaio, Fagani, W Arosi, Owa	—	—
NCV	Loh	Loh	—	—
SV	—	—	—	—
NCal	Nengone, Drehu, Iaai	—	—	—
Mic	Lamotrek	Marshallese (s.s.), Satawalese	Chuukese	—
Fij	Tavua	Macuata, Koroalau	—	—
Pn exc EPn	Tongan, Tokelau, E Futunan, Anuta, Tikopia, Pileni	—	Rennellese (s.s.), W Futunan, Takuu, Luangiua, Sikaiana	Pukapuka
EPn	Marquesan	—	Rapa, Maori	Tahitian, Hawaiian

Two groups in Table 2.6 deserve comment. The first of these is the Central Papuan subgroup of PT. Sinaugoro, Motu, Roro and North Mekeo are the Central Papuan languages in the sample. The Roro structure is shown in Figure 2.5. Sinaugoro, Motu and Roro retain bifurcate merging in the parental generation but have generational terminology in ego’s generation, i.e. all ego’s siblings and cousins are identically labelled, according to seniority. This was probably the situation in Proto Central Papuan. North Mekeo, however, adds generational terminology at G₊₁, shown in Figure 2.6. This is almost a full-blown generational terminology, as are the others listed in column 3 of Table 2.6.

The other group in need of comment is Polynesian. Only Ifira-Mele and West Futuna (= Futuna-Aniwa) are listed in Table 2.5 as bifurcate merging. There is good reason to infer that the structure of the Ifira-Mele terminology is the result of contact with NCV languages of Nguna (§2.3.3). The same may be true of West Futuna, which has had contact with the languages of Tanna (SV).(Lynch & Fakamuria 1994). It is probable that PPn had a terminology like that of Tongan and similar to Roro (Figure 2.5), in which all blood-relatives of ego’s generation were labelled in the same way as ego’s siblings, but the parental generation had dedicated terms for FZ and MB (Marck 1996). Loss of FZ and MB terms in various of the outliers and in PEPn resulted in structures like that in North Mekeo. However, Pukapuka, Tahitian and Hawaiian have also replaced PPn *tina- ‘mother’ (§2.4.1.2.1) and *tama- ‘father’ (§2.4.1.2.1) with a reflex of *matuqa- ‘parent’ (§2.4.1.2.3). The elimination of the sex distinction in the parental generation has resulted in a purely generational structure in Lowie’s (1928) sense, leading to the adoption of ‘Hawaiian’ as the term for a generational structure in Murdock’s (1949) classification of kinship terminologies.

3.2. Other structural features involving ego’s blood relatives⇫

3.2.1. EGO’s siblings and parallel cousins⇫

Bifurcate merging means that in almost every Oceanic terminology, EGO’s parallel cousins, i.e. EGO’s father’s brother’s and mother’s sister’s children, are labelled with the same terms as EGO’s siblings (§2.3.1). However, there is considerable variation in this labelling. For example, there may be a terminological distinction between s.s. and o.s. siblings/parallel cousins; or a seniority distinction among siblings/parallel cousins.¹⁴

These distinctions are structurally independent of bifurcate merging. The most frequently found distinction is between s.s. and o.s. siblings/parallel cousins. Following this pattern, speakers of earlier New Guinea Pidgin took the terms susa and brata from English sister and brother but used them in an Oceanic manner. If I was female, my brata was my sister (s.s. sibling) and my susa was my brother (o.s. sibling). These two categories are further subdivided in many Oceanic terminologies. A common pattern is illustrated by Nakanai (§2.2.3; Figure 2.3), where s.s. siblings are divided into those elder (tua-) or younger (tari-) than EGO, but o.s. siblings are divided into female (hatavile-) and male (hatamale-).

There are thus three sibling distinctions in Oceanic languages:

• relative sex: same sex as EGO vs opposite sex from EGO (sG vs oG)
• seniority: younger than EGO vs elder than EGO (yG vs eG)
• ALTER’s sex (Z vs B)

Eleven configurations of these distinctions occur in Oceanic languages, and Table 2.7 shows the number of times each configuration occurs in the sample.¹⁵ Column 1 shows whether a terminology makes the relative sex distinction (y[es] or n[o]). If it does, then there are entries in columns 2 to 5. Columns 2 and 4 show whether the seniority distinction is applied respectively to s.s. and o.s. siblings. Columns 3 and 5 do the same for ALTER’s sex. But if column 1 shows that a terminology lacks the relative sex distinction, then columns 4 and 5 are blank, because o.s. siblings are referred to by the same terms as s.s. siblings. For example, Nakanai makes the s.s./o.s. distinction (column 1), the seniority distinction for s.s. siblings (column 2), and uses different o.s. terms according to ALTER’s sex (column 5). It is thus one of the 17 languages with the configuration in row 6. The column totals show that only four terminologies distinguish ALTER’s sex in the terms for s.s. siblings, and that only three distinguish seniority of o.s. siblings. The four configurations that include one of these distinctions are shown in grey (3, 4, 7 and 8) and are presumably the outcomes of local accidents.¹⁶ Theoretically, 20 configurations are possible, but only 11 occur in Table 2.7—or more properly ten, as row 9 contains question marks for unknown values.¹⁷ Whilst one can speculate about the reasons for the absence or near-absence of certain configurations, it is clear that they reflect absences from the POc terminology (§2.3.5). Table 2.8 shows the languages in the six most attested categories in Table 2.7.

Table 2.7 Distributions of sibling distinctions across Oceanic languages

	1	2	3	4	5
config.	sG/oG	ysG/esG	sZ/sB	yoG/eoG	oZ/oB	Totals
1	y	n	n	n	n	42
2	y	n	n	n	y	9
3	y	n	n	y	n	2
4	y	n	y	n	n	3
5	y	y	n	n	n	38
6	y	y	n	n	y	17
7	y	y	n	y	n	1
8	y	y	y	n	y	1
9	y	y	n	?	?	8
10	n	n	n	—	—	5
11	n	y	n	—	—	12
Total y	121	77	4	3	27
Total n	17	61	134	110	86
					Grand total:	138

Table 2.8 Terminologies according to the better represented configurations in Table 2.7

	Config 1	Config 2	Config 5	Config 6	Config 10	Config 11
Adm	Sori, Lele		Pak	Baluan, Lou	Yapese	Mussau
SJ and NNG	Takia, Adzera	Mapos Buang, Lakurumau, Roviana, Marovo, Big Nambas, Anejom	Wogeo, Yabem, Middle Watut
PT	Dobu, Galea, Tubetube, Sudest		Kilivila, Muyuw, Kalauna, Bwaidoga, Sinaugoro			Motu, Roro, N Mekeo
MM	Vitu, W Kara, Nalik, Lihir, Notsi, Lamusong, Madak, Barok, Sursurunga, Nehan, Petats, Hahon, Mono-Alu	Lakurumau, Roviana, Marovo	Varisi, Nduke, Simbo	Nakanai
SES	Gae, Lengo, Sa’a, W Arosi, E Arosi		Bugotu, Gela, Birao, Longgu, To’aba’ita, Baegu		NWGR	Kwaio, Fagani, Owa
NCV	Loh, Araki, Nduindui, N Ambrym, Nguna	Big Nambas	Mota, NE Ambae, Raga, Vao, Ninde	Sinesip
SV		Anejom		Kwamera
NCal			Drehu, Iaai	Kumak, Ajië, Xârâcùù, Nengone	Belep, Jawe	Nyelâyu, Cèmuhî, Paicî
Mic	Satawalese, Lamotrek	Chuukese		Marshallese	Nauruan
Fij	Moala		most Fijian languages			Matailobau, Koroalau
Pn exc EPn	E Futunan, Anuta, Tikopia, Ifira-Mele, W Futunan, Takuu, Luangiua, Sikaiana	Tokelau, Rennellese	Pukapuka	Tongan, Pileni
EPn				Rapa, Maori, Tahitian, Marquesan, Hawaiian

In the vast majority of Oceanic languages, terms for parallel cousins are identical to those for siblings, but there is a dimension that is not represented here. In some languages, as one might expect, terms for EGO’s younger and elder siblings are applied to EGO’s younger and elder parallel cousins. In other languages, however, the terms make reference not to EGO’s seniority but to seniority of the cousin’s parent relative to EGO’s parent. For example, if the parallel cousin is EGO’s mother’s sister’s child, the choice of the ‘younger’ or ‘elder’ term depends on whether the cousin’s mother is younger or elder than EGO’s mother. This matter is not pursued here, as the data often do not allow us to determine who seniority refers to in a given terminology.

3.2.2. Children⇫

Strangely, perhaps, discussion of terms for children is necessarily preceded by a short return to terms for father’s sister and mother’s brother, a topic in the discussion of bifurcate merging (§2.3.1). Their relevance here is that ‘mother’s brother’ and ‘sister’s child’ are the same (reciprocal) term in some languages. In other scattered languages, ‘mother’s brother’ is the same term as ‘father’s sister’.¹⁸ However, these two categories almost never overlap: that is, there is no language in the sample in which ‘mother’s brother’, ‘father’s sister’ and ‘sister’s child’ are all the same term. Two exceptions are the Fijian dialects Moala and Lakeba, where FZ, MB and ♂ZC are all vuŋo-.

The role of mother’s brother and his relationship with his sister’s children are discussed briefly at the end of §2.2.3. In 52 terminologies each refers to the other by the same dedicated dedicated term (column 1 of Table 2.9). In 55 others, there is a separate term for each relationship (column 2). In yet others, there is no dedicated term for ‘sister’s child’ and the latter is not distinguished from other children. It is either referred to simply as a ‘child’ or by another kinship term, the narrow meaning of which is shown in parentheses (column 3). Among the items in column 2, in Simbo (MM), Pukapuka, Rapa and Māori (EPn) the MB term is the term for ‘father’.

An odd phenomenon, commented on several times in the literature, occurs in certain SES languages. In Bugotu, Gela, Lengo and Owa, the MB term reflects POc *tubu- ‘grandfather’ but does not retain this meaning. It is also used reciprocally for ♂ZC.

In a few WOc languages an apparent mirror-image of the MB/♂ZC relationship is enshrined in the terminology: FZ and ♀BC are referred to by the same term. These languages are Muyuw, Galea, and Motu (PT), and Vitu and Nakanai (MM). However, terms used only by a female ego and terms for children other than ego’s child and ego’s sister’s child are often not recorded, so we do not know how widespread this reciprocal usage is.

Table 2.9 Terms for male EGO’s sister’s child

	1	2	3
	MB = ♂ZC	MB, ♂ZC	MB, ♂ZC = C
Adm	Yapese, Mussau, Lele, Pak	Baluan, Lou	Sori (sG)
SJ and NNG	Sobei, Wogeo, Takia, Mapos Buang	Yabem, Bukawa, Middle Watut
PT	Kilivila, Muyuw, Kalauna, Dobu, Galea, Sinaugoro, Motu, Roro	Tubetube	Bwaidoga, N Mekeo
MM	Vitu, Nakanai, W Kara, Lakurumau, Nalik, Lamusong, Barok, Sursurunga, Trivett’s Tolai, Matupit Tolai, Nehan, Hahon, Mono-Alu, Nduke, Marovo	Patpatar, Petats, Simbo	Lihir (oGC), Notsi (oGC), Varisi (CC), Roviana
SES	NWGR, Birao, Longgu, To’aba’ita, Kwaio, Sa’a, Fagani, W Arosi, E Arosi
NCV	Loh , Araki, Vao	Mota, Raga, N Ambrym, Akei, NE Ambae, Nduindui, Big Nambas, Ninde, Sinesip, Nguna
SV		Sye, Kwamera	Anejom
NCal	Nyelâyu , Kumak, Fwai, Pwaamei, Wahmwaang, Cèmuhî	Yuanga, Bwatoo, Arhâ, Ajië, Paicî, Xârâcùù, Xaragure, Nengone, Drehu	Kwenyii
Mic	Nauruan	Marshallese, Satawalese, Lamotrek	Chuukese
Fij	Bauan	Vuda, Wayan, Nadrau, Tokatoka, Matailobau, Tavua, Deuba, Nakoroka, Cakaudrove, Moala, Lakeba, Lau	Koroalau (CC), Macuata
Pn exc EPn		Tongan, Tokelau, E Futunan, Anuta, Tikopia, Ifira-Mele, W Futunan
EPn	Marquesan	Pukapuka, Rapa, Maori	Tahitian, Hawaiian

In a scattering of languages the children of EGO’s male blood relatives and those of female blood relatives are referred to by different terms. In these languages the term for the children of females is usually the term used for ♂ZC. Our data are in this respect too incomplete for a listing of languages to be meaningful.

In a relatively small number of languages a sex distinction is made among children. There are dedicated kinship terms for female and male children in Pak, Baluan and Lou (Adm), Lihir, Notsi, Lamusong, Barok and Patpatar (MM, New Ireland), N Ambrym (NCV) and Tongan, Tokelau, E Futuna and Rennellese (Pn).

3.2.3. Grandparents and grandchildren⇫

Some Oceanic languages have a dedicated term for mother’s mother’s brother (e.g. Tolai kaku-; see also §2.3.3), but a majority encode all grandparents and members of EGO’s grandparents’ generation with a single term (often a reflex of POc *tubu-; §2.4.1.4.1). Other languages do make a distinction between female and male members of the grandparent generation. Our data are incomplete in this regard, but these languages include: Notsi, Sursurunga and Mono-Alu (MM); Bugotu, Sa’a, W Arosi and Fagani (SES); Akei, Big Nambas, Ninde and Nguna (NCV); almost all NCal languages (exceptions are Belep in the north, Kwenyii in the south, and the languages of the Loyalties); Marshallese, and Lamotrek (Mic), numerous and scattered Fijian languages, and Ifira-Mele (Pn).

Owa (SES) has a general term pʷāpʷā for members of the grandparent generation as well as wauwa- ‘grandfather’ and tura- ‘grandmother’. It is quite possible that there are other languages with a similar configuration that do not show up in the sample. Sori and Pak (Adm) are recorded with separate terms for PF, MM, FM. On the other hand, a few languages have no ‘grandparent’ terms, instead using ‘mother’ and ‘father’ for both parents and grandparents. They are Simbo, Nduke and Roviana (MM), Numèè (NCal) and Satawalese and Chuukese (Mic).

Data on ‘grandchild’ terms are more complete, and fall into two categories:

• a ‘grandparent’ term is used reciprocally of both grandparents and grandchildren;
• there is a dedicated ‘grandchild’ term.

These categories mostly follow genealogical boundaries. A ‘grandparent’ term is used reciprocally to refer to grandchildren in Mussau, NNG, PT, MM, SES and Marshallese (Mic). A dedicated grandchild term is used in Yapese, Adm, NCal including the Loyalties, and throughout Fiji and Polynesia. NCV languages in the sample are divided. Most have a ‘grandchild’ term, but the northern languages Loh, Mota and Araki use ‘grandparent’ reciprocally.

Languages that also use ‘mother’ and ‘father’ for grandparents are exceptions to these generalisations. Roviana uses ‘father’ reciprocally for a grandchild, Nduke has ‘child’s child’, and Numèè (along with Kwenyii), Satawalese and Chuukese use ‘child’ for grandchildren as well as children.

3.3. Generational skewing⇫

A small number of languages in the sample have ‘cross’ terms with extended meanings that are mostly quite unlike those in Nakanai. These are exemplified by NE Ambae, a north Vanuatu language (Figure 2.7).¹⁹

Within the merged portion of the terminology, the structure of NE Ambae (Figure 2.7) differs from Nakanai (Figure 2.3) only in the unusual addition of tue- ‘s.s. sibling’ as a cover term for both tehi- ‘younger s.s. sibling’ and toka- ‘elder s.s. sibling’.

It is the ‘cross’ portion of the NE Ambae terminology in Figure 2.7 that is of particular interest. On the left-hand side of the figure, father’s sister is also called ratahi- ‘mother’, and so are the female descendants in her matriline, i.e. FZD, FZDD, FZDDD. Their brothers (FZS, FZDS, FZDDS) are all called tama- ‘father’. The children of each tama- are labelled in the same way as the children of the three tama- in the merged portion of the figure.

The right-hand ‘cross’ portion of the terminology is different. The children of EGO’s mother’s brother are called natu- ‘child’, and their children vaɣabui- ‘grandchild’.

The technical term for this phenomenon is ‘generational skewing’. On the left of the figure, ‘mother’, ‘father’ and the two sibling terms are used of women a generation or more below EGO’s mother. On the right of the figure, ‘child’ and ‘grandchild’ are used of individuals a generation above EGO’s child and grandchild. Skewing is also found in various locations outside the Austronesian family (see Lounsbury 1964a and the contributions to Trautmann & Whiteley 2012).

There is a plentiful literature on generationally skewed terminologies, asking how they come into being and what function they serve. However, a few things seem reasonably clear. There are two varieties of generational skewing. One is like NE Ambae, in Murdock’s terms a ‘Crow’ terminology. Its mirror-image, ‘Omaha’, is not represented in Oceanic. Kohler (1897) found that ‘Crow’ terminologies were correlated with matrilineal societies, ‘Omaha’ with patrilineal. This correlation stands as a high probability, but not an absolute, as examples of patrilineal Crow and matrilineal Omaha have since been found. The converse does not hold, i.e. unilineal (matri- or patrilineal) descent is not a good predictor of generational skewing. For example, Nakanai society is matrilineal but its terminology is not skewed.

The witness of speakers of languages with a skewed terminology gives us some indication of how skewing happens. Deacon’s ‘best’ informant for Semiang (= Sinesip, NCV) volunteered the information that if a male EGO’s mother’s brother died, by custom EGO would marry the MB’s widow, thereby becoming ‘father’ to MB’s children, and calling them ‘child’, accounting for the right-hand side of Figure 2.7. Conversely, if EGO is among the children of the deceased, then their cross-cousin becomes their ‘father’, accounting for the left-hand side (Deacon 1934:77-78). It seems that these uses in this circumstance have become entrenched, resulting in the skewed terminology.

Fortune’s (1932) description of Dobu (PT) kinship is significant in two ways. The Dobu terminology has two states, A and B in Figure 2.8.. State A resembles the Nakanai structure of Figure 2.3. There are dedicated terms for mother’s brother and father’s sister, and a single term for cross-cousins. After father’s death, state B becomes the norm, with generational skewing on the father’s side. FZD inherits her mother’s label, and her husband receives the label ‘father’. Fortune (1932:37–38) provides the reason for this. The father’s place in the matriline is taken by his heir. The heir is the father’s sister’s son, so he assumes the title ‘father’, and his sister becomes ‘father’s sister’. Fortune’s consultants’ explanation of the generationally skewed state B closely resembles that of Deacon’s Sinesip consultant. As their locations are 1950 km apart as the crow flies, they are clearly independent pieces of evidence as to the reason for generational skewing.

This explains how generational skewing arises. But the fact that two terminological states coexist(ed) in Dobu society also explains how Oceanic comes to have terminologies like state A alongside terminologies like state B, not to mention terminologies with fossil fragments of state B (see below). No other Oceanic terminologies with two states are known to us, but we infer that they must exist or have existed in order for the transition between the two states to take place.

Kronenfeld (2012) describes a two-state terminology in the Fanti language of Ghana. The situation is similar to that in Dobu. Kronenfeld takes issue with the concept of a ‘Crow’ terminology, viewing it as a generally skewed overlay on an existing bifurcate merging (‘Iroquois’) terminology. The Fanti two-state terminology can be taken as corroborating Fortune’s account of Dobu and its historical role.

To ‘qualify’ as a fully skewed system the changes illustrated by NE Ambae in Figure 2.7 are expected:

1. on the father’s side
   1. father’s sister’s matrilineal female descendants (FZD, FZDD, FZDDD) are referred to with the same term as FZ;
   2. father’s sister’s matrilineal male descendants (FZS, FZDS, FZDDS) are called ‘father’;
   3. the children of father’s sister’s matrilineal male descendants (FZSC, FZDSC, FZDDSC) are called by the same terms as EGO’s siblings (i.e. as the children of a ‘father’)
2. on the mother’s side
   1. mother’s brother’s child is called by the term for EGO’s child’;
   2. mother’s brother’s grandchild is called by the term for EGO’s grandchild.

Twenty-one languages in the sample show signs of generational skewing, but only three, Mota, NE Ambae and Nguna (all NCV) qualify fully. The evidence is set out schematically in Table 2.10.²⁰ The abbreviations in the first row are relationships in a genealogical tree. The abbreviations beneath them represent the narrow sense of the term that is used for that relationship. In NE Ambae, for example, the relationship of father’s sister (FZ) is referred to by the term for mother (M), the relationship of father’s sister’s son (FZS) by the term for father (F), and so on. Terms that do not reflect generational skewing are omitted for readability’s sake.

Table 2.10 Oceanic languages reflecting generational skewing

gloss	FZ	FZS	FZD	FZSC	FZDS	FZDD	MBS	MBD	MBSC	MBDC
Yapese	M	F	M	G	G		C	C		CC
Sori	FZ	F	FZ					C
Lele	FZ¹	F	FZ			FZ		D
Pak	FZ	FZS FZ		G	FZS	FZ		D	CC
Kilivila	PP	F	PP		FZ	FZ	C	C
Dobu state B	FZ	F	FZ
Sudest	FZ						C²	C²
Nakanai				PP	PP				PP	PP
Notsi				PP	*PP ³				PP	*PP³
Lamusong	FZ			G	FZ				G	FZ
Usen Barok				G					G
Trivett’s Tolai	PP		PP			PP		F		PP
Mota	M	F	M	G	F	M	C	C	PP	PP
Akei	M	F	M				C	C	CC
NE Ambae M		F	M	G	F	M	C	C	CC	CC
Nduindui	M	F	M	G	G		C	C	CC	CC
Raga	FZ	F	FZ		F	FZ	C	C		CC
Ninde	FZ	F	FZ	G	G		C			CC
Sinesip	M	F	M				C	C
Nguna	FZ	F	FZ	G	F	FZ	C	C	G	G
Ifira-Mele	FZ	F	FZ				C	C

• ¹ Father’s mother is also called by the FZ term.
• ² Alternates with the term for cross-cousin.
• ³ The term is tubo-, reflecting POc *tubu PP, not reflected as PP in Notsi.

The backbone of the skewed system is the terms for the matrilineal female, FZ, FZD and FZDD, shown in grey. The observation here is that whatever term is used for FZ is also used for FZD, FZDD and FZDDD (FZDDD is not shown; cf Figure 2.7). In some languages, e.g. Lele (Adm), the same term is projected upwards, i.e. used for FM. The matrilineal female’s brother (FZS, FZDS) is labelled as ‘father’ (F). In Pak (Adm) a special term is used for FZS, which is then also used for FZDS.

Abbreviations in italics represent terms that have evidently been copied from elsewhere in the paradigm. For example, Nduindui terms are almost identical to those of its neighbour, NE Ambae, but the term used for FZDS is not the expected term ‘father’ but ‘sibling’, copied from FZSC. In Nduindui matrilineal moieties have been replaced by patrilineal sections. The Nduindui terms for father’s sister’s children retain the old nomenclature, but it is lost in the following generations (Allen 1964a).

Nakanai and Notsi (both MM) show only a fragment of an assumed formerly skewed system. In a terminology with skewing on the mother’s side, mother’s brother’s children (MBS and MBD) are called ‘child’ (C), and their children (MBSC and MBDC) are called ‘grandchild’. The latter are, however, of the same generation as EGO’s children. Nakanai and Notsi retain only this peculiarity of a skewed system, and have copied the ‘grandchild’ term to the children of the paternal cross-cousins.

The sample languages include two Tolai dialects, Matupit Tolai and an unlocated dialect described by Trevitt (1940). This pair is interesting in that Matupit Tolai is unskewed, whereas ‘Trevitt’s Tolai’ (labelled thus in Table 2.10) shows signs of skewing. Both dialects are shown in Figure 2.9. Assuming the hypothesis that unskewed and skewed versions of a terminology initially coexist, Matupit Tolai represents the unskewed version, Trevitt’s Tolai a frozen skewed version. Trevitt’s Tolai uses the term tubu- ‘grandparent’ for FZ, FZD and FZDD, presumably because, working upward, FM is tubu-.²¹ FZS and FZDS are not called by the expected term ‘father’, but instead nauva- FZS is extended to FZDS. A peculiarity of Trevitt’s Tolai is that it uses tama- reciprocally of ‘father’ and ‘child’, and this is then predictably extended to mother’s brother’s daughter (but not son), whose daughter is then labelled tubu- ‘grandchild’.

The only Pn language with skewing is Ifira-Mele, and we attribute this to contact with neighbouring Nguna on the island of Efate.

3.4. The geography of typology⇫

A review of Tables 2.5 to 2.10 reveals that typology rarely follows geographic or genealogical boundaries closely. This is scarcely surprising, as changes in terminological structure are likely to occur independently in different places. Without the terms themselves, changes do not provide a secure basis for determining shared innovations, as some of the comments published with Marshall (1984) explain. But this does not mean that the typological investigation above is irrelevant to the reconstruction of POc and the subsequent history of Oceanic languages.

The distribution of birfurcate merging (Table 2.5) is significant. It is strong in the Admiralties, in the New Britain/New Ireland area of MM, in New Caledonia other than the Loyalties, and in most of Fiji. This distribution, coupled with the fact that it is easier to derive other structures from bifurcate merging than vice versa, suggests that the POc terminology had bifurcate merging. However, it was lost among the small islands of Micronesia and eastern Polynesia.

The data supporting generational skewing are admittedly found in only a small minority of Oceanic languages, but their presence in quite widely distributed languages of the Admiralties, PT, the New Britain/New Ireland area of MM, and NCV points to the presence of skewing in POc.

The relabelling of father’s sister as ‘mother’ was seemingly a shared innovation in Proto NW Solomonic (MM). It is an areal tendency in SES (Table 2.5). On the other hand the naming of all siblings and cousins, including cross-cousins, by the same terms seems to be a broad tendency that operates sporadically everywhere except SV and mainland New Caledonia. Change towards a generational structure went furthest in eastern Polynesia (Table 2.6).

Table 2.11 Reconstructed POc kinship terms

Gloss	Reference terms		Address terms
‘father, father’s brother, mother’s sister’s husband’	*tama-	*ama	*mama
‘mother, mother’s sister, father’s brother’s wife’	*tina-, *kina-	*ina	*(ñ,n)ana, *nai
‘mother’s brother’	*matuqa-		*wawa
‘child of ego, s.s. sibling or parallel cousin’	*natu-, *tu-
‘♂sister’s child’	*[qa]lawa
‘♂sister’s child’ (?)	*pa(s,c)u, *pa(s,c)ua-
‘kin of grandparent and grandchild generations’	*tubu-	*ubu	*bubu, *[bu]bui
	*tabu-	*abu[a]
	*wawa[-]		*tete
‘kin of grandchild generation’	*makubu-
‘great-great-grandparent, great-great-grandchild’	*bawa[-]		*sese-
‘younger s.s. sibling, younger s.s. parallel cousin’	*taci-, *kaci-
‘elder s.s. sibling, elder s.s. parallel cousin’	*tua-, *tuaka-		*kaka
‘♂sister’, ♂female parallel cousin’	*papine-
‘♀brother’, ♀male parallel cousin’	*mʷaqane-
‘friend, companion; relative of ego’s generation’	*tuRaŋ
‘spouse’	*qasawa-
‘woman, wife’	*pine
‘s.s. sibling-in-law’	*ipaR/*ipa-
‘o.s. sibling-in-law, younger than ego’	*taci-
‘o.s. sibling-in-law, older than ego’	*tua-
‘parent- or child-in-law’	*rawa

3.5. Reconstructing the Proto Oceanic kinship terminology⇫

The POc kinship terms reconstructed in the subsections of §2.4.1 (blood relatives) and §2.4.2 (in-laws) are listed in Table 2.11 and plotted as a tree diagram in Figure 2. First, POc kinship terms in §2.4 were reconstructed. Then the terminology was assembled from the terms and their reconstructed glosses, but with an eye on the typology of terminologies presented in §2.3. On the division of terms into reference and address terms, and the division of certain reference terms into *t-initial and *t-less, see §2.4.1.1.

The glosses of the terms for ‘father’ and ‘mother’, based on glosses of their reflexes, support the typological evidence above that POc terminology was bifurcate merging (§2.3.1). The s.s. sibling terms distinguished *taci- ‘younger s.s. sibling’ from *tua- or *tuaka- ‘elder s.s. sibling’. The o.s. sibling terms differentiated between the directly possessed terms *papine- ‘♂o.s. (female) sibling’ and *mʷaqane- ‘♀o.s. (male) sibling’. These were also the POc terms for ‘wife’ and ‘husband’ (vol.5:50–55), but when they referred to spouses they were indirectly possessed. Putative *lopu- ‘o.s. sibling’ is omitted from both the figure and the table as its POc status is uncertain (§2.4.1.5.3).

There is broad agreement that early Oceanic society was matrilineal. In his survey of earlier findings of Oceanic matrilineality Blust (1981c:70–72) notes that Codrington (1891:30), Deacon (1934:705), and Capell and Lester (1945-46, part 4:315) had all suggested that at least in some parts of Melanesia, where societies with moieties are interspersed with societies that lack them, moieties are a retention, a view that Blust endorses. In his 1994 paper he shows that POc matrilineality was inherited from PMP, and points to Murdock’s (1968a) survey of sibling terminologies, which finds that the same-sex/opposite-sex opposition is correlated with the presence of descent groups and especially with matrilineality. More recently Hage (1999; this volume), Hage & Marck (2003) and Marck (2008) have also argued vigorously in favour of POc matrilineality, partly on genetic grounds.

If POc was indeed matrilineal, then it presumably had two moieties like those of Tolai, NE Ambae and a number of other Oceanic languages (§2.2.3). Symptomatic of this is the presence of two terms for ‘child’. POc *[qa]lawa- denoted in its narrow sense ‘sister’s child’ (i.e. the one to whom EGO was mother’s brother) and in its classificatory sense ‘female parallel cousin’s child’. The default term was *natu-, the only term that, as in Tolai, switches moiety according to EGO’s sex: opposite moiety to EGO if EGO is male, same moiety as EGO if EGO is female.

The existence of a dedicated term, *matuqa, for EGO’s mother’s brother (literally ‘the old one’) reflects his role among his kin (§2.2.3) and is further evidence of matrineality. If the society were patrilineal, the leading male among his relatives would be his father.

3.6. Generational skewing in Proto Oceanic⇫

Previous reconstructions of POc kinship terms seem to have assumed that there must have been a separate term for EGO’s father’s sister. Perhaps there was, but one cannot reconstruct it (§2.4.1.2.6). Given that most other terms for blood relatives are easily reconstructable, this blank in the terminology demands an explanation, and a fairly straightforward one emerges from the discussion of generational skewing in §2.3.3. Like Dobu, POc apparently had two states, a state A in which there were dedicated terms for father’s sister and for cross-cousins and their children, and a state B in which these terms were replaced by generationally skewed terms. Assuming that early Oceanic speakers lived in small hamlets with perhaps one or two extended families scattered along coastlines (vol.1:62), it is possible that children rarely heard the state A terms, rendering them liable to the replacement which is visible in the data. The few generationally skewed terminologies that exist today reflect a freezing of all or parts of state B (Table 2.10, plus others not included in the sample), but most Oceanic terminologies reflect state A, with variations resulting from a past two-state situation.

In the POc terminology in Figure 2 the slots with terms in square brackets are essentially gaps. No distinct terms can be reconstructed for them. They are in the lineages that involve cross-cousins: father’s sister and her descendants and mother’s brother’s descendants. The terms in the brackets are the generationally skewed terms that would have been used in state B, with the exception of ‘father’s sister’. This is shown as [*FZ] because the POc term is unknown. It was perhaps *tina- ‘mother’ or *tubu- ‘grandparent’, both discussed below.

The evidence of the languages listed in Table 2.10, like NE Ambae (Figure 2.7) and Trevitt’s Tolai (Figure 2.9), leads us to expect that EGO’s mothers’ brother’s children will be called ‘child’, despite the fact that they are of EGO’s generation. This extract from the data for ‘child’ in §2.4.1.3.1 confirms this. Glosses not relevant to generational skewing have been removed. A majority of the reflexes are from NCV languages. This is not surprising, as generational skewing, at least in the north of the region, is mentioned in the ethnographic literature. The presence of reflexes from the Admiralties, PT, MM (western Solomons), the Loyalties and Micronesia makes it likely that the sense ‘mother’s brother’s child’ is of POc antiquity.

POc		*natu-	‘child, s.s. sibling’s child, parallel cousin’s child, mother’s brother’s child’
Adm	Pak	naro-	[♀MBC] ‘child’
PT	Kilivila	latu-	[♂MBC] ‘child’
MM	Tolai	tama-	[MBC] ‘father, child’ (Trevitt 1940)
MM	Marovo	tu-	[MBC] ‘child’
NCV	Mota	natu-	[MBC] ‘child’ (Codrington 1891)
NCV	Raga	nitu-	[MBC] ‘child’
NCV	Akei	natu-	[BC MBC] ‘child’
NCV	Daakaka	natɛ-	[MBC] ‘♂child’ (Deacon 1927; von Prince 2012: 112-117)
NCV	Labo	nitu-	[♂BC ♂MBC] ‘child’ (Deacon 1934: 91–96)
NCV	Naman	netu-	[MBC] ‘child’ (Deacon 1934: 104–106)
NCV	Uripiv	natu-	[MBC] ‘child’ (Deacon 1934: 124)
NCV	Neve’ei	natu-	[MBC] ‘child’
NCV	Nguna	natu	[MBC] ‘child’ (Facey 1989)
NCal	Iaai	noko-	[MBC] ‘child’ (Ray 1917)
Mic	Puluwatese	nawɨ-	[MBC] ‘child’

One would also expect that the children of EGO’s MBC would be called by the same term as ‘grandchild’, and the data give some support to this notion. POc apparently used two terms for ‘grandchild’. POc *tubu- ‘grandparent’ was used reciprocally of ‘grandchild’. POc *makubu- was not reciprocal and meant ‘grandchild; kin two generations below ego’. The terms in the set below reflect a variety of ‘grandchild’ terms, but have in common the fact that they are also used for mother’s brother’s grandchild, who was of EGO’s child’s generation, not EGO’s grandchild’s.

POc		*tubu-	‘grandparent, grandchild’
POc		*makubu-	‘grandchild; kin two generations below ego’
Yap	Yapese	tuŋin	[BCC MBDC] ‘grandchild’
Adm	Pak	makapu-	[sGCC MBSC] ‘grandchild’
MM	Tolai	tubu-	[FZ FZD FZDD MBDS] ‘grandchild’ (Trevitt 1940)
MM	Babatana	bazu-	[♂MBDC] ‘grandchild’
NCV	Akei	mabʷi-	[BCC MBSC] ‘grandchild’
NCV	Mota	tupu-	[MBCC FZCC] ‘grandparent, grandchild’ (Codrington 1891)
NCV	Naman	nābu-	[MBSS] ‘grandchild’
NCV	Neve’ei	nābu-	[MBSS] ‘grandchild’
NCV	Nguna	sūli	[BCC ♀MBCC] ‘grandchild’
NCV	Labo	neiʙü-	[MBDS ♂MBSS] ‘grandchild’
Fij	Lautoka	viaŋo-	[WZC MBCC] ‘grandchild’
Fij	Tavuki	makubu	[BCC MBCC] ‘grandchild’

NE Ambae and Trevitt’s Tolai also suggest that father’s mother, father’s sister (FZ), FZD, FZDD and so on, i.e. father’s maternal line, were either labelled ‘mother’ or ‘grandparent’ in POc. The evidence for this is less than the set above, but the strength of the ‘mother’ set below (from §2.4.1.2.2) is reinforced by the Nduindui and Chuukese glosses, which include both FZD and FZDD. The POc term was presumably the well supported *tina- ‘mother’, despite the fact that Lele, Mota and Nduindui all have another term for ‘mother’. The evidence for POc *tubu- in this role is thinner but the glosses are convincing.

Yap	Yapese	ci-tini(ŋo)-	[FBW FZ FZD]	‘mother, mother’s sister’
Adm	Lele	nane-	[FZD]	‘mother, mother’s sister’
NCV	Mota	veve-	[FBW FZD]	‘mother, mother’s sister’(Codrington 1891)
NCV	Nduindui	re-tahi-	[PZ FZD FZDD]	‘mother’
NCV	Akei	tina-	[PZ FZD]	‘mother’
Mic	Puluwatese	yīn	[PZ FZD]	‘mother’
Mic	Chuukese	yine-y	[PZ PM FZD FZDD]	‘mother’
PT	Kilivila	tabu-	[FZ FZD FZDD]	‘grandparent’(Lounsbury 1965; Malinowski 1929)
MM	Tolai	tubu-	[FZ FZD FZDD MBDS]	‘grandparent’(Trevitt 1940)

In NE Ambae the brother of the father’s lineal female is called *tama- ‘father’, i.e. FMB, F (and FB), FZS, FZDS and FZDDS are all *tama-. Again there is respectable evidence for this among the data in §2.4.1.2.1

POc		*tama-	‘father, father’s mother, mother’s sister’s husband, father’s sister’s son, father’s sister’s daughter’s son, father’s sister’s daughter’s daughter’s son’
Yap	Yapese	ci-timo(ŋo)-	[FB MZH FZS] ‘father’
Adm	Pak	dramo-, tamo-	[FB ♀FZS] ‘father’ (Mead 1934: 347–349)
Adm	Bipi	tama	[FB ♀FZS] ‘father’ (Mead 1934: 350–351)
PT	Kilivila	tama-	[FB PZH FZS] ‘father’ (Lounsbury 1965; Malinowski 1929)
NCV	Mota	tama-	[FB MZH FZS] ‘father’
NCV	Lombaha	tama-	[FB FZS] ‘father’
NCV	Longana	tama-	[FB FZS] ‘father’
NCV	Akei	tama-	[FB FBS FZS] ‘father’
NCV	Daakaka	timya-	[FB FZS] ‘father’
NCV	Labo	tama-	[FB MZH FZS] ‘father’
NCV	Nguna	mama	[MZH FZS FZDS FZDDS] ‘father’ (Facey 1989)
Mic	Chuukese	seme-	[PB FZS FZDS] ‘father’

Collectively, the evidence above is respectable support for the hypothesis that the gaps in the POc tree are there because POc has a stage B. Some of the bracketed terms in Figure 2 are reconstructed not by the comparative method but by local inference. For example, if mother’s brother’s child was labelled *natu-, otherwise ‘child’, then mother’s brother’s child’s child was probably labelled *makubu-, otherwise ‘grandchild’.

4. Reconstructing Proto Oceanic kinship terms⇫

In this section POc terms are presented with supporting evidence. Among other things, we were interested in ascertaining whether a reconstructed term was a POc innovation or an inheritance from an earlier Austronesian interstage. As in other chapters of these volumes, we often rely on others’ reconstructions of PAn, PMP or PEMP terms as evidence of inheritance. Where a relevant reconstruction has not been made, however, we have explored SHWNG languages to see whether a term is reconstructable to PEMP—and therefore inherited into POc—and have cited relevant SHWNG data. SHWNG languages fall into four subgroups: South Halmahera (SH), Raja Ampat (RA), Bomberai (Bom) and Cenderawasih Bay (CB).

Some PPn kin terms are shown with two forms, like *tina- and *tinana ‘mother’ (Pawley 1967:262–263; Wilson 1982:35–60, 96-99). The first term of the pair, like other items ending in a hyphen, retained POc direct possession of kin terms whereby the possessor suffix was added directly to the stem (vol.1:32; vol.5:75–76), but this occurred only with a singular possessor.²² With a non-singular possessor and optionally with a singular possessor an indirect possession construction was used, e.g. PPn *t-o-tā tama-na [ART-PCL-P:1PL father-P:3SG] ‘our father’ (Wilson 1982:97). In the indirect construction the possessed noun retained a fossil reflex of the POc *-ña P:3SG, giving the second term of the pair.

There are several pitfalls associated with the glossing of kinship terms. At one end of the continuum, a term is glossed with a single term, e.g. ‘father’, yet, given that Oceanic kinship terminologies are classificatory, it is almost certain that a ‘father’ term has wider reference that includes ‘father’s brother’ and ‘mother’s sister’s husband’. However, defining the ‘father’ term as denoting F, FB and MZH in many cases does not cover the term’s actual denotation, which may well be potentially boundless (see §2.2.3). A few publications, e.g. Fingleton (1986) on Tolai, offer definitions of this kind. The usual ethnographic gloss, however, is a long string of relatives, like Elbert & Monberg’s (1965:13) gloss of Rennellese tama-/tamana: F, FB, FZH, FFBDH, FFBS, FMZS, MH, FMZDH, MZH, MMZDH, HF, WF. A string of this kind is sometimes difficult to interpret, as it omits a relationship that is implied if the term denotes a boundless class. Does this mean that the authors have omitted a possible gloss? We cannot tell, as ethnographers do not always explicate the structure of the kinship system, but instead focus on the behaviours of various kin towards each other. In any case, the interpretation of a gloss list like the one for Rennellese ‘father’ is dependent on an understanding of the structure of the kinship system.

Often when the long-list glosses in a cognate set are compared, we find that perhaps only the first three items on the list agree across cognates. Beyond these it is impossible to draw any conclusion about the larger category denoted by the reconstruction.

In light of these considerations and to save space, long lists of glosses are generally not given in the definitions below. These can sometimes be found in the ethnographies listed in the appendix to this chapter. Instead, we restrict ourselves to the most straightforward glosses, like ‘father, father’s brother,’ FZH. In the interests of readability the first gloss (or two) is given in plain English and subsequent glosses as abbreviations. On occasion, a definition of the whole category is provided, like the one above (‘Father or any male blood relative…’).

To avoid too much clutter in the cognate sets below, we have not given the source of most cognates. As with other chapters and volumes, the lexical source(s) for most languages are listed in Appendix A. The present chapter also draws on the ethnographic sources listed at the end of the chapter. The source of a cognate is given only where (a) we have consulted two or more works for that language; (b) the source is itself a comparative work dealing with a number of languages; or (c) only one item in the chapter is drawn from that source.

In some Oceanic languages, there are two terms for the same ALTER, a reference term used in talking about that ALTER, e.g. ‘my father’ or ‘your child’, and an address (or vocative) term used to indicate whom one is talking to, e.g. ‘Dad, may I go outside?’ Address terms are marked with the abbreviation ADDR. However, there are languages in which this distinction is not made, i.e. one says, ‘My father, may I go outside?’ Diachronically, it is quite common to find a reference term that reflects an early address term.

4.1. Blood relatives (consanguineal kin)⇫

4.1.1. Origins⇫

Table 2.12 POc variant terms for close blood relatives

	1	2	3	4
‘father’	*ama	*tama-	(*kama)	*mama	(§2.4.1.2.1)
‘mother’	*ina	*tina-	*kina	*nana	(§2.4.1.2.2)
	*ubu	*tubu-	*kubu
‘grandparent, grandchild’	—	*tibu-	—	*bubu	(§2.4.1.4.1)
	*abu	*tabu-	—
‘younger s.s. sibling’	*aci	*taci-	*kaci	—	(§2.4.1.5.1)
‘elder s.s. sibling’	—	*tua-, *tuaka-	—	*kaka	(§2.4.1.5.2)

Table 2.13 Distribution by Oceanic subgroups/regions of POc variant terms for close blood relatives

column:	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
SHWNG	y	y	y	y	y	y		y	y	y	y	y		y	y	y
POc	*ama	*tama-	*kama	*mama	*ina	*tina	*kina	*ñaña	*ubu	*kubu-	*tubu-	*bubu	*tibu-	*abu	*tabu-	*aci	*taci-	*kasi-	*tua-	*tuaka-	*kaka
Mussau		y					y				y						y		y
Adm		y			y	y		y			y	y		y	y		y
NNG	y	y	y	y	y	y	y	y	y		y	y	y	y	y	y	y		y	y	y
PT	y	y	y	y	y	y				y	y	y					y		y	y
Kb		y			y	y	y	y			y			y			y		y	y
MM		y		y	y	y		y			y	y	y				y	y	y	y
NWS		y		y		y			y		y	y	y			y	y	y		y	y
SES		y		y		y		y									y	y	y	y	y
NCV		y		y	y	y		y	y		y	y	y	y	y		y		y	y
SV		y		y		y					y						y		y
NCal		y		y		y		y			y			y			y		y	y
Mic		y				y					y						y
Fij		y				y		y			y	y					y		y	y
Pn		y				y		y	y		y	y					y			y

‘y’ means ‘yes, there are reflexes of the item at the head of the column’.

POc forms denoting certain close blood relatives appear to form a (defective) paradigm, shown in Table 2.12. The distribution of their reflexes across Oceanic subgroups is shown in Table 2.13. The most widely reflected forms are those with *t- (column 2 of Table 2.12). They were evidently reference forms and usually took a possessor suffix (§2.2). The forms in column 4 were address forms, apart from *kaka, which has a long history as both an address and a reference form. In some Oceanic languages reflexes of the other items in column 4 are also reference forms, and it is quite likely that this extension of function had already begun in POc. However, it is impossible to identify a functional difference between the forms in column 2 and the less widely reflected forms in column 1 or the rarely reflected forms in column 3 (Table 2.13).²³

The relationships among the forms in Table 2.12 are obvious, but gaps and irregularities, absences of functional contrasts between forms, and three rows of ‘grandparent/grandchild’ forms instead of one all suggest that this is not entirely a paradigm in the usual morphological sense, but in part a set of relationships (re)created by analogical innovations.

This hypothesis is reinforced when the POc reconstructions are compared with their PMP ancestors in Table 2.14. POc forms in *k- (column 3 of Table 2.12) are missing from Table 2.14, because no PMP forms in *k- are reconstructed. The only *k- form for which there is non-Oceanic evidence is *kama ‘father’, and this can be reconstructed only to PEMP.²⁴ On the other hand, Blust (1979, ACD) shows that ‘basic’ forms like *ama and *t- forms like *tama did form a paradigm in PMP, resulting from the prefixation of PAn *ta- to basic forms, the latter becoming POc reference terms. Blust (1979) also proposes that PMP created address forms by shifting stress to the final syllable of the basic form. Hence, e.g., PMP *áma (reference) vs *amá (address).²⁵ He suggests that this led to initial-vowel loss, giving *ma and reduplicated *mama, the PMP address term reflected in Oceanic.

Blust’s reduplication hypothesis certainly appears to explain *mama, *nana and *bubu, but it doesn’t work for *kaka (and Blust doesn’t suggest that it does). The latter is of PAn antiquity, yet the corresponding reference form *tuaka is apparently a POc innovation, so *kaka cannot be derived from it.

Table 2.14 PMP and POc variant terms for close blood relatives

	PMP	POc	PMP	POc	PMP	POc
‘father’	*ama	*ama	*tama	*tama-	*mama	*mama
‘mother’	*ina	*ina	*tina-	*tina-	*nana ¹	*nana
	*umpu	*ubu	*tumpu	*tubu-
‘grandparent, grandchild’	(*impu)	—	(*timpu)	*tibu-	*bubu	*bubu
	*ampu	*abu	— ²	*tabu-
‘younger s.s. sibling’	*huaji, *haji ³	(*aci)	*tuaji, *taji ³	*taci-	—	—
‘elder s.s. sibling’	*aka	—	(*teka)	*tua-, *tuaka	*kaka	*kaka

Note: Parentheses indicate that a form is very weakly attested, i.e. 2-3 WMP reflexes in ACD. Bolding indicates that a PAn form can be reconstructed.

• ¹ ACD: Proto Philippine *nana ‘term of address for older female relative’.
• ² *tampu is reconstructed only as far back as PEMP (ACD).
• ³ Under the ACD entry for PAn *Suaji is a note thaty irregular reflexes leave open the possibility that we should reconstruct both *Suaji and *Saji’. The reconstructions shown here follow from this.

The origin of *tuaka is puzzling. It consists of *tua- and *-ka, and seems to be an analogical creation to fill the previously empty slot in column 2 of Table 2.12 with a *t- form. The second element, *-ka, is an obvious abbreviation of *kaka. The first element, *tua-, perhaps already existed. It is reflected in many Oceanic languages with a possessor suffix. Its origin is not clear. Milke’s (1958b) reconstruction of *tuqaka with *-q- implies he thought *tuqa- was POc *tuqa ‘old’ (vol.2:204–205; vol.5:68; see §2.4.1.5.2). He had toyed with this thought in his 1938 paper, but in 1958b, with more data and better reconstructions, he explicitly denies it in a footnote, but leaves *-q- in place.²⁶

An alternative hypothesis—that *tuaka reflects PMP *teka and its POc continuation *toka, both in the ACD—is beset with difficulties. First, PMP *teka is supported by only one non-Oceanic reflex (Bintulu təka ‘elder sibling’) and, as noted in the ACD entry, no corresponding basic PMP form †*eka can be reconstructed. The ACD lists several reflexes of putative POc *toka (see §2.4.1.5.2 for data). However, deriving *tuaka from *toka implies an otherwise unknown change *-o- > -ua-. As the converse change, *-ua- > -o-, vowel coalescence in unstressed syllables, is far more natural, the many apparent reflexes of putative *toka can more economically be explained as reflexes of POc *tuaka, and this is how they are treated in §2.4.1.5.2).

The use of ‘appears to’, ’seems to’ and ‘apparently’ in the foregoing paragraphs indicates that we are dangerously close to speculation, and the evidence points only to probable conclusions. Pages could be spent speculating about matters raised by Tables 2.13 and 2.14, but we restrict ourselves here to just one further question: Why are there apparently three rows for ‘grandparent, grandchild’?

According to the ACD, PMP had basic forms with the three vowels *u-, *i- and *a- (Table 2.12).²⁷ However, the distribution of reflexes of the *i- row reconstructions suggests that the latter are an artefact of the comparative method. There are no known Oceanic reflexes of PMP *impu. Table 2.13 shows that Oceanic forms apparently reflecting POc *tibu- only occur in subgroups where POc *tubu- is also reflected. This suggests that these forms are actually reflexes of POc *tubu- in which *-u- has become -i- in the unstressed first syllable, e.g. *tubú-gu > *tibú-gu ‘my grandparent’. For this reason, in §2.4.1.4.1 POc *tubu- and *tibu- are treated as a single cognate set. We lack the data to comment on non-Oceanic forms, but note that the PMP forms in the *i- row are poorly supported in the ACD, *impu with two WMP and one CMP reflex and *timpu with one western MP and the Oceanic reflexes mentioned above.

This leaves a contrast between the forms in the *u- row and the *a- row. There is no functional distinction between them, and both are well attested in Oceanic (Table 2.13).

Table 2.14 shows that PMP *bubu is of PAn antiquity and had a different labial consonant from the *-mp- of forms in the other columns of the ‘grandparent/grandchild’ rows. However, if Blust’s reduplication hypothesis applies here, it will have generated POc forms indistinguishable from those descended from PMP *bubu.

4.1.2. Parent and child generations⇫

4.1.2.1. Father, father’s brother

The three terms for ‘father’ shown in Table 2.14, *tama-, *ama and *mama, are presented first. These are followed by a further term, *tata, that is either an address term or derived from one. There are numerous reflexes of an apparent *papa in the data, but it is impossible to determine in many cases whether these are loans from European colonial languages or not, and they are omitted here.

The POc reference term for ‘father’ was *tama-. Many glosses reflect the fact that POc kinship terminology was bifurcate merging (see §2.3.1), i.e. terms for ‘father’ also denoted father’s brothers (FB). A common third gloss is ‘mother’s sister’s husband’ (MZH). In a bifurcate merging terminology a mother’s sister is also a ‘mother’, and the husband of any ‘mother’ is a ‘father’.

PAn		*ta-amax	‘father’ (ACD)
PMP		*t-ama	‘father, father’s brother’ (Blust 1980a)
POc		*tama-	‘father, father’s brother, mother’s sister’s husband’ (ACD; Milke 1958b)
Yap	Yapese	ci-timo(ŋo)-	[FB MZH FZS] ‘father’
Adm	Mussau	tama-	[MZH] ‘father, father’s brother’
Adm	Baluan	tama-	[FFBS] ‘father, father’s brother’
Adm	Lele	tam-	‘father, father’s brother’
Adm	Drehet	teme	[MZH PGC FGSC ♀ZH] ‘father, father’s brother’
SJ	Sobei	tema-	‘father, father’s younger brother’
NNG	Tuam	tama-	‘father, father’s brother’
NNG	Mangap	tama-	‘father, father’s brother’
NNG	Lukep	tama-	‘father’
NNG	Mengen	tama-	‘father’
NNG	Takia	tama-	[FFBS MZH etc] ‘father, father’s brother’
NNG	Wogeo	tama-	[PZH] ‘father, father’s brother’
NNG	Manam	tama-	[EF FBS] ‘father, father’s brother’
NNG	Yabem	tama-	‘father’
NNG	Adzera	rama-	[MZH] ‘father, father’s brother’
NNG	Mapos Buang	ama-	[GF] ‘father, father’s brother’
NNG	Patep	ma	‘father, father’s brother’
PT	Sudest	rama-	‘father’
PT	Kilivila	tama-	[FB PZH FZS] ‘father’ (Lounsbury 1965; Malinowski 1929)
PT	Dobu	tama-	[etc MZH etc] ‘father, father’s brother’
PT	Iamalele	tama	[MZH] ‘father, father’s brother’
PT	Sinaugoro	tama-	[MZH] ‘father, father’s brother’
PT	West Mekeo	ama	[ZH] ‘father, parent’s brother’
MM	Vitu	tama-	[MH] ‘father, parent’s brother’
MM	Nakanai	tama-	[HB MZH FEB MEZH] ‘father, father’s brother’
MM	East Kara	tama-	‘father’
MM	Lamasong	tama-	[MZH FFF MMF] ‘father, father’s brother’
MM	Tolai	tama-	[MBC] ‘father, father’s brother’ (Fingleton 1986)
MM	Nehan	tama-	[etc] ‘father, father’s brother’ (Nachman 1978)
MM	Hahon	tama-	[FZH] ‘father, father’s brother’
MM	Banoni	tama-	[F{PsG}S MZH] ‘father, father’s brother’
MM	Varisi	tama-	[FZH MZH] ‘father, father’s brother’
MM	Marovo	tama-	[FZH FZS] ‘father, father’s brother’
MM	Maringe	kma-	‘father, father’s brother’
SES	Bugotu	tama-	[PB] ‘father’ (Bogesi 1948)
SES	To’aba’ita	(θa)ama-	‘father’
SES	Kwara’ae	ama-	[MZH] ‘father, father’s brother’ (nearly obsolete)
SES	Sa’a	ama-	[FZH] ‘father’
SES	Bauro	(w)ama	[etc] ‘father, father’s brother’ (wa- MASC)
PNCV		*tama-	‘father, father’s brother’ (Clark 2009)
NCV	Mota	tama-	[FB MZH FZS] ‘father, father’s brother’ (Needham 1960; Vienne 1984)
NCV	Lombaha	tama-	[FB FZS] ‘father, father’s brother’ (Allen 1964a)
NCV	Raga	tama-	[FZS FZDS ZDH ♀DH] ‘father’
NCV	Akei	tama-	[FB FBS FZS] ‘father, father’s brother’ (Guiart 1958)
NCV	Araki	rann̼a-	[MZH] ‘father, father’s brother’
NCV	Southeast Ambrym	tamo	‘father’ (Lynch 1996)
NCV	Neve’ei	teme-	[FB FZS] ‘father’
NCV	Namakir	tama-	‘father’
PSV		*e-tme-	‘father, father’s brother’ (Lynch 2001c)
SV	Sye	e-tme-	‘father, father’s brother’
SV	Lenakel	rəmə-	‘father, father’s brother’
SV	Anejom̃	e-tma-	[FB FZH] ‘father’
PNCal		*tama-	‘father, father’s brother’ (Ozanne-Rivierre 2000)
NCal	Belep	cama-	[MZH] ‘father, father’s brother’
NCal	Nêlêmwa	kã̄mã	‘father, father’s brother’
NCal	Drubea	tẽ̄-	‘father, father’s brother’
NCal	Iaai	kame-	‘father, father’s brother’
PMic		*tama	‘father, father’s brother’ (Hage and Marck 2002)
Mic	Kiribati	tama	[♂EF] ‘father, father’s brother’ (Lambert 1981)
Mic	Marshallese	cema-	[♂EF ♂EFB] ‘father, father’s brother’ (Spoehr 1949a)
Mic	Puluwatese	hām, hæmæ-	[FZS] ‘father, father’s brother’
Mic	Woleaian	tama-	‘father’ (Hage and Marck 2002)
Mic	Chuukese	seme-	[PB FZS FZDS] ‘father, parent’s brother’ (Goodenough 1951)
Fij	Wayan	tama-	[MZH] ‘father, father’s brother’
Fij	Bauan	tama-	‘father, father’s brother’
Fij	Moala	tama-	[P{PsG}S FMZS MH WMB] ‘father, father’s brother’
PPn		*tama-, *tama-na	‘father, father’s brother’ (Marck 1996; Pawley 1981: 284)
Pn	Tongan	tama(i)	‘father, father’s brother’ (Völkel 2015)
Pn	Samoan	tamā	[etc] ‘father, father’s brother’ (Holmes 1957)
Pn	Rennellese	tama-, tamana	[FZH FFBDH F{PsG}S MH MZH EF] ‘father, father’s brother’
Pn	Pileni	tamo	[MZH] ‘father, father’s brother’

Reflexes of *ama ‘father, father’s brother’ are few and restricted to PNGOc. It is only the presence of non-Oceanic reflexes that allows us to reconstruct this term to POc.

PAn		*amax	‘father’ (ACD)
PMP		*ama	‘father, father’s brother’ (Blust 1994)
POc		*ama	‘father, father’s brother’
NNG	Tuam	ama	‘father’
NNG	Tami	ama(ʔ)	‘father’
NNG	Medebur	ama	‘father’ (Z’graggen 1974a)
NNG	Labu	ama	‘father’
PT	Tawala	ama-	‘father’

POc *mama either was an address term or was descended from one, as the ACD gloss notes. Chowning (1991) also thinks that POc *mama was an address term, as it has this function all along the north coast of Papua New Guinea.

It seems that the denotation of *mama changed in Proto New Caledonian to ‘elder s.s. sibling’. Indeed, the change may have been earlier, as this is also its meaning in Neve’ei. This appears to be an extension of seniority status to elder s.s. siblings which also affects *tata below.

PAn		*mamah	‘father, father’s brother’ (ACD: ’father__s brother’)
PMP		*mama	‘father, father’s brother’ (ACD: ’a young child__s term of address for his father, vocative of *ama, parent__s younger sibling, junior uncle’)
POc		*mama	‘father, father’s brother’ (ACD; Chowning 1991)
SJ	Sobei	mam	‘father, father-in-law’ (ADDR) (Sterner 1992)
NNG	Gitua	mama	‘father’
NNG	Sio	mama	‘father’
NNG	Mindiri	mamə	‘father’
NNG	Wogeo	mam	‘father, father’s brother’
NNG	Kairiru	mam	‘father’
NNG	Yabem	mama	‘father’
NNG	Numbami	mama	[MZH] ‘father, father’s brother’
NNG	Middle Watut	mama	‘father’
PT	Gapapaiwa	mama(i)	‘father’
PT	Western Motu	mama	[MZH] ‘father, father’s brother’ (Lister-Turner & Clark 1954)
MM	Lavongai	mama(i)	‘father’ (Fast and Fast 1989)
MM	Nalik	mama-	‘father’ (Volker 2020)
MM	Usen Barok	mama	‘father, father’s brother’
MM	Patpatar	məma-	[MZH FZ] ‘father, father’s brother’
MM	Varisi	mamae	[FZH MZH] ‘father, father’s brother’
MM	Roviana	mama	‘father’ (Waterhouse 1949)
MM	Kokota	mama	[PZH] ‘father, father’s brother’
SES	Bugotu	mama	[PB] ‘father’ (ADDR) (Bogesi 1948)
SES	Lengo	mama	‘father, father’s brother’
SES	Longgu	mama	[PZH] ‘father, father’s brother’
SES	Baegu	mā	‘father, father’s brother’
SES	Kwaio	maʔa	‘father, father’s brother’
SES	Sa’a	mamaʔa, ma’a	[FZH] ‘father’
PNCV		*mama	‘father’ (Clark 2009; Proto North Vanuatu *mama)
NCV	Mota	mama	‘father’ (REF and ADDR; § 1882:67)
NCV	Nduindui	mama	[FZS BS FZH FZDH] ‘father, father’s brother’
NCV	Neve’ei	mameh	[♂{FeB}S {WeZ}♀H] ‘♂elder brother’
NCV	Nguna	mama	[MZH FZS FZDS FZDDS] ‘father’ (Facey 1989)
NCal	Fwâi	māma	‘♂elder brother’
NCal	Kwênyii	mõmʷã	‘elder s.s. sibling’
NCal	Nengone	mama	‘elder s.s. sibling’ (Ozanne-Rivierre 2000)
Mic	Mokilese	mʷāmʷa	‘father’

A more widely distributed POc address term is *tata, used for anyone called ‘father’ (F, FB, MZH etc) and by extension for other senior males, primarily elder s.s. siblings and any mother’s brother. In some languages this extension has become entrenched. In Notsi, for example, the term denotes both ‘mother’s brother’ and his reciprocal, ‘♂sister’s child’.

POc		*tata	‘address term for any male called “father’’ and for other senior males’ (ADDR)
NNG	Kove	tata	‘father, father’s brother’ (ADDR) (Chowning 2009)
NNG	Sio	tata	‘elder s.s. sibling’
PT	Iamalele	tata	‘elder s.s. sibling’
MM	Vitu	tata	[FG MH MB (ADDR)] ‘father’
MM	Notsi	tata	[♂ZC ♂MMZS ♂WFB ♂MMBDH ♀HFZS ♀HZH] ‘mother’s brother’
MM	Lamasong	tata	[♂ZC MMMB ♂ZDDC ♂{PsG}DC] ‘mother’s brother’
MM	Siar	tata	‘father’ (Frowein 2011)
MM	Uruava	tatá-	‘elder s.s. sibling’
NCV	Raga	tata	[FZS FZDS ZDH ♀DH] ‘father’
NCV	Araki	ta	[MZH] ‘father, father’s brother’
NCV	Apma	tata	‘father’
NCV	North Ambrym	tāta, teta	[♂S BS FFF FFBS MFZDS MFMB MH MZH] ‘father, father’s brother’ (Franjieh 2012)
NCV	Paamese (South)	tātā	‘father’
NCV	Big Nambas	tatei	‘father, father’s brother’
NCV	Neve’ei	tate	‘father, father’s sister’s son’
NCV	South Efate	tata	‘father’ (Lynch 2004d)
SV	Ura	dera	‘father, father’s brother’ (ADDR)
SV	Lenakel	tata	‘father, father’s brother’ (ADDR)
NCal	Belep	caca	[MZH] ‘father, father’s brother’
NCal	Fwâi	tāra, tāru	‘father, father’s brother’
NCal	Dehu	kakā	‘father, father’s brother’
Mic	Carolinian	tāta	‘father, father’s brother’
Fij	Tavuki	tata	‘father, father’s brother’
Pn	Samoan	tā	‘father, father’s brother’ (Milner 1966)
Pn	Aniwa	tata	[MZH] ‘father, father’s brother’ (Guiart 1961)

4.1.2.2. Mother, mother’s sister

The first four terms below are those shown in Table 2.14: POc *tina-, *ina, *kina- and *(ñ,n)ana. These are followed by two address terms, POc *nai and *nene, and finally PPn *faqe(e).

Lynch (1996) provides commentary on the fact that POc *tina- is more frequently replaced by another lexical item than POc *tama- ‘father’. This is especially so in EOc languages.

PAn		*ta-ina	‘mother’ (ACD)
PMP		*tina	‘mother, mother’s sister’ (Blust 1980a; ACD)
POc		*tina	‘mother, mother’s sister’ (Milke 1958b: ‘mother’)
Adm	Baluan	tina-	[MFBD etc] ‘mother, mother’s sister’
Adm	Pak	hi-rino-	[PZ FZD FZDD FFBD FFBDD] ‘mother’
Adm	Bipi	tine	[♀MFZSD ♂ZW] ‘mother, mother’s sister’
Adm	Nyindrou	tine-	‘mother, mother’s sister’
SJ	Sobei	tina-	‘mother, mother’s younger sister’
NNG	Tuam	tina-	‘mother’
NNG	Sio	tino-	[etc] ‘mother, mother’s sister’ (Groves 1934b)
NNG	Lukep	tina-	‘mother’
NNG	Aria	tna-	[F{PsG}D] ‘mother, mother’s sister’
NNG	Takia	tna-	[etc] ‘mother, mother’s sister’
NNG	Wogeo	tina-	[PZ PBW WZ ♂BW] ‘mother’
NNG	Yabem	tɪna-	‘mother’
NNG	Adzera	rina-	[FBW] ‘mother, mother’s sister’
PT	Sudest	təna-	‘mother’
PT	Iamalele	ina	‘mother’
PT	Sinaugoro	sina-	[FBW] ‘mother, mother’s sister’
PT	West Mekeo	ina-	[PZ BW] ‘mother’
MM	Vitu	tina-	‘mother’
MM	Nalik	dina	‘mother’
MM	Patpatar	e-tna-	‘mother’
MM	Tolai	tina-	‘mother, mother’s sister’ (Trevitt 1940)
MM	Nehan	tina-	‘mother’ (Glennon and Glennon 2006)
MM	Hahon	cina-	[PZ FBW] ‘mother’
MM	Banoni	cinna	[M{PsG}D FW FBW] ‘mother, mother’s sister’
MM	Varisi	sina-	[MBW FZ FBW] ‘mother, mother’s sister’
MM	Marovo	tina-	[BW] ‘mother, mother’s sister’
SES	Gela	tina-	[FZ] ‘mother, mother’s sister’
SES	Lengo	tina-	‘mother’
SES	To’aba’ita	θa-ina-	‘mother’
SES	Arosi	ina-	[etc] ‘mother, mother’s sister’
SES	Bauro	ka-ina-	‘mother’ (ka- FEM)
PNCV		*tina	‘mother, mother’s sister’ (Clark 2009)
PNCV		*ra-tina-	‘mother, mother’s sister’
NCV	Vurës	re-tna-	‘mother’
NCV	Nokuku	tina-	‘mother’
NCV	Araki	ra-rna-	[FBW] ‘mother, mother’s sister’
NCV	Lonwolwol	i-rña-	‘father’s sister’
NCV	Paamese (South)	la-tino-	‘mother’
NCV	Neve’ei	sena-	[FZSW] ‘mother, mother’s sister’
PSV		*ri-(t,c)inV-	‘mother, mother’s sister’ (Lynch 2001c)
SV	Sye	n-rin(me)-	‘mother, mother’s sister’
SV	Kwamera	ri-nn̥u-	[FW FBW EFZ FFZD] ‘mother, mother’s sister’
SV	Anejom̃	ri-si-	[MBW] ‘mother, mother’s sister’
PNCal		*tina-	‘mother, mother’s sister’ (Ozanne-Rivierre 2000)
NCal	Nemi	tne-	‘mother, mother’s sister’
NCal	Fwâi	tʰẽ-	‘mother, mother’s sister’
NCal	Drubea	ñā	‘mother, mother’s sister’
NCal	Iaai	hiñe	‘mother, mother’s sister’
PMic		*tina	‘mother, mother’s sister’ (Bender et al. 2003a)
Mic	Kiribati	tina-	‘mother’ (Groves et al. 1985)
Mic	Marshallese	cine-	[MBW ♂EM ♂EMZ ♀HZ] ‘mother, mother’s sister’ (Spoehr 1949a)
Mic	Puluwatese	yīn, yina-	[PZ FZD] ‘mother’ (Elbert 1974)
Mic	Woleaian	sila-	‘mother’ (Hage and Marck 2002)
Fij	Tokatoka	tina-	[FZ M{PsG}oG] ‘mother’
Fij	Bauan	tina-	‘mother, mother’s sister’
Fij	Moala	tina-	[M{PsG}D MM{PsG}DD MFF{PsG}D F{PsG}D HFZ HF{PsG}D F{PsG}SW FF{PsG}SSW] ‘mother, mother’s sister’
PPn		*tina-, *tinana
Pn	Samoan	tinā-	‘mother’ (Milner 1966)
Pn	Rennellese	tina-u, tinana	[PZ MBW P{PsG}D F{PsG}SW FBW EM] ‘mother’
Pn	Pileni	hina-	‘mother’
Pn	Rarotongan	tīnana	‘mother’

Reflexes of *ina are somewhat thinly spread, but their distribution indicates that the term was present in POc. It is not clear how this might have differed from *tina-.

PAn		*ina	‘mother, mother’s sister’ (ACD)
PMP		*ina	‘mother, mother’s sister’ (ACD)
POc		*ina	‘mother, mother’s sister’
Adm	Titan	ina	‘mother’
NNG	Sengseng	ina	‘mother’ (ADDR) (Chowning 1991)
PT	Bwaidoga	ina-	[FZ] ‘mother, mother’s sister’
PT	Western Motu	ina	[FBW] ‘mother, mother’s sister’ (Lister-Turner & Clark 1954)
MM	Maututu	ila	‘mother’ (Goodenough 1997)
MM	Meramera	ina	‘mother’ (Goodenough 1997)
MM	West Kara	ina	‘mother’
SES	Gela	ino	[PZ] ‘mother’
PNCV		*ina	‘mother, mother’s sister’
NCV	North Ambrym	ina	[BD MFM {PsG}SD FFFZ BSSD ZSDD MFZDD {PoG}ZDDD MBW FFMBW ZDSW SSD EMM ♂D ♂DSSW ♂SWBW ♀SDD ♀DSW HZD] ‘father’s sister’ (Franjieh 2012:226, 371)
NCV	Port Sandwich	ina	‘mother’
SV	Lenakel	inə-	‘mother, mother’s sister’

Only three Oceanic reflexes of *kina- occur in the available data, but their distribution points to their presence in POc. Again, there is no obvious functional difference from *tina-.

POc		*kina-	‘mother’
Adm	Mussau	kina-	[FBW] ‘mother, mother’s sister’
NNG	Rauto	kina	‘mother’ (Ross, fieldnotes)
MM	Bola	kina-	‘mother’

Chowning (1991) reconstructs *(ñ,n)ana as an address term. Although few of our examples are explicitly glossed as addressed terms, she is very probably right. Several Fijian and Pn languages reflect *nau ‘mother’, an apparent Proto Central Pacific innovation

PMP		*nana	‘mother’ (ACD: Proto Philippine *nana ‘term of address for older female relative’)
POc		*ñaña	‘mother, mother’s sister’ (Chowning 1991)
Adm	Lele	nane-	[FZD] ‘mother, mother’s sister’
NNG	Gitua	nana	‘mother’
NNG	Sio	nana	‘mother’
NNG	Megiar	naŋ	‘mother’ (Z’graggen 1974a)
NNG	Bam	nan	‘mother’ (Z’graggen 1974b)
NNG	Ali	ñan	‘mother’ (Z’graggen 1974b)
MM	Bola	nana	‘mother’ (Goodenough 1997)
MM	East Kara	nina-	‘mother’
MM	Nalik	nana	‘mother’ (Volker 1998)
MM	Sursurunga	nana	‘mother’
MM	Siar	nana	[MB] ‘mother’ (Frowein 2011)
MM	Hahon	na	[PZ FBW] ‘mother’
MM	Banoni	nana	‘mother’
MM	Kubokota	ñaña	‘mother’
SES	Tolo	nana-	‘mother, mother’s sister’
SES	Owa	nana	‘mother’
NCV	Neve’ei	nane, na-	[FZSW] ‘mother, mother’s sister’
NCal	Nêlêmwa	ñaña, ñāña	‘mother, mother’s sister’
NCal	Fwâi	ñaña	‘mother, mother’s sister’
NCal	Kwênyii	ñã̄	‘mother, mother’s sister’
NCal	Dehu	nenæ	[(ADDR)] ‘mother, mother’s sister’
NCal	Nengone	ne-	[(ADDR)] ‘mother’ (Ray 1917)
Mic	Carolinian	nāna	‘mother’
PCP		*nana, *nau	‘mother, mother’s sister’ (Chowning 1991)
Fij	Vuda	nana	‘mother’
Fij	Tavuki	nau	‘mother, mother’s sister’
Fij	Tokatoka	nana	‘mother’
Fij	Bauan	nana	‘mother’ (Capell 1941)
Fij	Moala	nana, na, nau	[etc] ‘mother’
Pn	Anutan	nau	[SW] ‘mother-in-law’
Pn	Tikopia	nana, nau	‘mother’
Pn	Pileni	nana	‘mother’

Blust (1979) reconstructs a set of PMP address terms in *-y, continued as POc *-i, whence POc *inai. Normally, PMP *-ay becomes POc *-e (Table 1.4 in §1.8.1), but this is apparently a case where an often-used address term does not undergo a regular phonological change (cf de Carvalho 2020).

PMP		*ina-y	‘mother (ADDR)’ (Blust 1979)
PEMP		*nai	‘mother (ADDR)’
CB	Warembori	nai	‘mother’
POc		*nai	‘mother (ADDR)’
NNG	Dami	nai	‘mother’
NNG	Takia	nei	[FBW] ‘mother, mother’s sister’
PT	Sudest	noi	[FBW] ‘mother, mother’s sister’ (Lepowsky 1981)
MM	Babatana	nai	‘mother’

PEMP/POc *nene below may reflect reduplication(s) of *nai above.

PEMP		*nene	‘mother’
SH	Sawai	nene	‘mother’ (Whisler 1996)
RA	Ambel	nén	[FBW] ‘mother, mother’s sister’
POc		*nene	‘mother’
NNG	Mindiri	nen	‘mother’
NNG	Kaiep	nen	‘mother’
SJ	Sobei	nen	‘mother’
MM	Usen Barok	nene-	‘mother, mother’s sister’
NCal	Kwênyii	nènè-	‘mother, mother’s sister’
NCal	Nengone	nene	‘mother, mother’s sister’ (Ozanne-Rivierre 2000)
Fij	Nadi	nēnē	‘mother’

In a number of forms in the *tina- and *(ñ,n)ana cognate sets above the term for ‘mother’ is preceded by a proclitic or prefix that reflects putative POc *drV- (Lynch 1996).²⁸ These forms are repeated together below. In Lavongai and Tigak, at least, this morpheme is still productive, and precedes any noun that denotes a mother, e.g. Lavongai ri təsi-m [HON parallel.sibling-P:2SG] ‘your sister’ (speaker is female, and her sister is a mother) (Fast 1990). Beaumont (1979) shows that Tigak ri- also precedes a mother’s proper name. Fast and Beaumont both gloss it as a prefix denoting respect for a mother. Lynch (1996) notes, on the other hand, that the SV reflexes are only found on the noun ‘mother’.

Adm	Pak	hi-rino-	[PZ FZD FZDD FFBD FFBDD]	‘mother’
MM	Lavongai	ri-nnə		‘mother’(Fast 1990)
MM	Tigak	ri-na-		‘mother’
NCV	Vurës	re-tna-		‘mother’
NCV	Lolsiwoi	ri-si-		‘mother’(Lynch 1996)
NCV	Araki	ra-rna-	[FBW]	‘mother, mother’s sister’
NCV	Paamese (South)	la-tino-		‘mother’
SV	Kwamera	ri-nn̥u-	[FW FBW EFZ FFZD]	‘mother, mother’s sister’
SV	Anejom̃	ri-si-	[MBW]	‘mother, mother’s sister’

Reflexes of this same morpheme are also in north Vanuatu, prefixed to other terms referring to a mother.

NCV	Mota	ra-veve-	[FZD FBW]	‘mother, mother’s sister’
NCV	Mota	r-asoa-		‘wife (who is a mother)’(Vienne 1984)
NCV	Merei	ra-ᵐbui-		‘mother’ (ADDR)(Lynch 1996)
NCV	Tolomako	ra-veve-		‘mother’(Lynch 1996)
NCV	Baetora	ra-ve		‘mother’(Lynch 1996)
NCV	Lombaha	ra-tahi-		‘mother, mother’s sister’
NCV	Lolovoli	re-tahi-		‘mother, mother’s sister’
NCV	Raga	ra-tahi	[MMM ZDDD]	‘mother’

Presumably the three last examples, from the Lombaha and Lolovoli dialects of NE Ambae and from Raga, use reflexes of POc *taci- ‘younger s.s. sibling’ originally made reference to a mother’s sibling.

In a number of Pn languages *tina- is replaced by a reflex of PPn *faqe(e). The second list of terms below, extracted from POLLEX, suggests that it may derived from PPn *faqele ‘woman who is pregnant or has recently given birth’.

PPn		*faqe(e)	‘mother, mother’s sister’
Pn	Tongan	faʔē	‘mother, mother’s sister’ (Völkel 2015)
Pn	Tongan	faʔē taŋata	‘mother’s brother’ (lit. ‘male mother’)
Pn	East Uvean	faʔe	‘mother, MZ’
Pn	Anutan	pae	[MPGD] ‘mother, mother’s sister’
Pn	Māori	ϕaea	‘mother’

PPn		*faqele	‘woman who is pregnant or has recently given birth’
Pn	Samoan	failele	‘mother with newborn child’
Pn	East Futunan	faʔele-ʔele	‘woman who has had first child’
Pn	Māori	ϕaere-ere	‘mother of one’s children, dam’

4.1.2.3. Parent

As far as we can tell, POc did not have a dedicated term for ‘parent’. PPn *matuqa meant ‘parent’ but, as a number of the glosses below show, it also retained its earlier sense ‘old, mature (of a person)’ (vol.4:68). There is just one clue that reflexes of *matuqa may have meant ‘parent’ earlier than PPn. This is Gapapaiwa (PT) *madua ‘mother’.

Further below are reflexes of the distinct PPn form *mātuqa ‘parents’.

PPn		*matuqa	‘parent’ (Marck 1996)
Proto Tongic		*motuqa	‘parent, parent’s sibling; old, mature (of a person)’
Pn	Tongan	motuʔa	‘parent; old (of people); old person’
Pn	Niuean	matua	‘parent, uncles and aunts’ (apparently a non-Tongic loan)
PNPn		*matuqa	‘parent, parent’s sibling; old, mature (of a person)’ (Marck 1996)
Pn	Samoan	matua	‘parent, old person’ (Williamson 1924, vol 2)
Pn	East Futunan	matua	‘parent’ (sometimes); ‘old’ (generally)
Pn	Rennellese	matuʔa	‘husband’
Pn	Tikopia	matua	‘parent’29
Pn	Pukapukan	matua	‘parent; adult’
PEPn		*matuqa	‘parent, parent’s sibling’ (Marck 1996)
Pn	Rapanui	matuʔa	‘father, mother, parents’ (POLLEX)
Pn	Māori	matua	‘parent; parent’s sibling’
Pn	Tahitian	metua	‘parent; parent’s sibling’
Pn	Marquesan	motua	‘father, father’s brother, father’s male cousin’
Pn	Hawaiian	makua	‘parent’

PPn *mātuqa ‘parents’ was the plural form of *matuqa ‘parent’, as widespread reflexes demonstrate. In some languages *mātuqa has lost the plural feature and become ‘parent[s]’, presumably to distinguish the ‘parent’ meaning from other meanings of *matuqa: Tuvalu and Nukuoro ‘old’, Rennellese ‘husband’.

PPn		*mātuqa	‘parents’ (Marck 1996; Marck 1996)
Pn	Tongan	mātuʔa	‘parents; old people’ (Völkel 2015)
PPn		*mātuqa	‘parents’ (Marck 1996; Marck 1996)
Pn	Samoan	ātua	‘parents’
Pn	Tokelauan	mātua	‘parents, mother, mother’s sister, older sister any senior female kin’ (Huntsman 1971)
Pn	Tuvalu	mātua	[PZ PPGC] ‘mother’
Pn	East Uvean	mātuʔa	‘parents’
Pn	Rennellese	mātuʔa	‘parent, spouse’s parent’
Pn	Anutan	mātua	‘parents’
Pn	Takuu	mātua	‘parent; clan head’
Pn	Luangiua	mākua	‘parents’
Pn	Māori	mātua	‘parents, relatives of parent’s generation’

EPn languages have replaced dedicated terms for ‘father’ and ‘mother’ with a reflex of POc *matuqa plus a term meaning ‘male’ or ‘female’, as shown here.

Pn	Rapa	metūa vahine		‘mother’
Pn	Rapa	metūa tāne	[FB]	‘father’
Pn	Tahitian	metua vahine		‘mother’
Pn	Tahitian	metua tane		‘father’
Pn	Hawaiian	makua-hine		‘mother’
Pn	Hawaiian	makua-kāne		‘father’

Niue and Pukapuka appear to have acquired their terms from an EPn languages. The Niue term matua is EPn in form, not Tongic.

Pn	Niuean	matua fifine	‘mother’
Pn	Niuean	matua tāne	‘father’
Pn	Pukapukan	matua wawine	‘mother’
Pn	Pukapukan	matua tāne	‘father’

4.1.2.4. Parent’s sibling

A number of apparent reflexes of PEMP/POc *kaka ‘elder s.s. sibling (ADDR)’ (§2.4.1.5.2) are used of a father’s or mother’s sibling. It is not clear whether there was a PEMP or POc term with this meaning, or whether an extension of the term to a parent’s sibling occurred independently in various languages.

Curiously, no SHWNG language reflects *kaka ‘elder s.s. sibling’, but both *kaka terms are reflected in Oceanic languages, although not in the same language.

The listing of definitions of the Kwamera term below from Lindstrom (1981) has been somewhat abbreviated. Lindstrom (1986) defines it as ‘person of first ascending generation, opposite marital moiety: uncle, aunt’.

PEMP		*ŋkaŋka	‘parent’s sibling’ (?)
RA	Ambel	kak	‘mother’s brother’
POc		*kaka	‘parent’s sibling’ (?)
MM	Lavongai	kakai	‘mother’s brother’ (Fast and Fast 1989)
MM	Sursurunga	kəkə-	‘father, father’s brother’
SES	Tolo	kaka	‘father, father’s brother’
SES	Kwara’ae	ʔaʔai	[♀BC MBW ♀HZC] ‘father’s sister’
SV	Kwamera	kaka	[{PoG}E P{PoG}SC EP GEP] ‘parent’s o.s. sibling’

4.1.2.5. Mother’s brother

Because the role of the mother’s eldest living brother is an important one in many matrilineal Oceanic-speaking communities (§2.2.3), and sometimes in communities that are no longer matrilineal, there is often a dedicated term for ‘mother’s brother’ (MB). Sometimes this term is also used for ♂sister’s son (♂ZS) or ♂sister’s child (♂ZC). That is, the term is used reciprocally.

Sometimes, the MB term is the one used by a man of his wife’s father (WF) or by any EGO of her/his spouse’s father (EF) (Blust 1980a, 1994). This seems odd until one recognises that in some communities—and evidently many more in the past—a man’s preferred marriage partner was his female cross-cousin, i.e. his MB’s daughter, so that after marriage his MB was his father-in-law (WF). This is/was the arrangement when cross-cousin marriage was asymmetric. In other communities, cross-cousin marriage was symmetric. One of the preferred marriage partners for both a woman or a man was a MB’s child, so after marriage the MB was also the father-in-law (EF).

On occasion, MB seems to be or to have been grouped with other relations, so that in various locations the term for MB is the same as the term for some other relationship(s). These terms are listed as members of cognate sets in the relevant subsections.

In a number of SE Solomonic languages the term for MB and ♂ZC reflects POc *tubu- ‘grandparent, grandchild’ (§2.4.1.4.1). In Lele (Adm) the *tubu- reflex is used of MB, MF and MMB; in Malalamai (NNG) of MB; and in Drubea (NCal) of MB and EF. Since this term is often used for all members of the grandparent generation, it seems that MB has been promoted by a generation.

In a few languages a reflex of *tama- or *tata ‘father’ (§2.4.1.2.1) is applied to all blood relations of father’s generation, including MB. Instances of *tama- used in this way are Iduna and W Mekeo (both PT) and of *tata Vitu (MM). There are also languages where a *tata reflex specifically denotes MB. In Notsi and Lamusong (MM) it denotes MB and ♂ZC, in Vurës (NCV) MB and FZH.

In Chuukic (Mic) languages, reflexes of POc *mʷaqane ‘♀brother’ have lost the o.s. sibling sense, but are combined with the term for ‘big’ in a compound literally meaning ‘big brother (of female)’ but actually denoting MB (§2.4.1.5.3). Ali (NNG) məne-, Kwara’ae (SES) wai and Tavuki (Fij) ŋane, all ‘mother’s brother’, are also reflexes of *mʷaqane, and one may speculate that they too once co-occurred with a term for ‘big’. Conversely the terms for MB in two languages of the New Ireland region, Lavongai lavə and Ramoaaina ləba, both reflect POc *lapʷat ‘big’, and perhaps reflect a similar compound but with loss of the *mʷaqane element.

Although few, there are also MB terms that seem once to have more literally meant ‘brother of mother’. In the Loyalty Islands (NCal) we find Iaai ma-hiñe- and Drehu mā-θin, both ‘MB’ (Ozanne-Rivierre 2000), where hiñe- and θin respectively mean ‘mother’. PPn *tuqa-tina is presumably derived from POc *tuqa ‘mature, full-grown, ripe, old’ (vol.5:67–68)³⁰ and *tina- ‘mother’ (§2.4.1.2.2). Note that PPn *tuqa- here is not the same element as *tua- (< POc *tuRaŋ) in PPn o.s. sibling terms (§2.4.1.5.3). It is, however, the root of PPn *ma-tuqa below.

PPn		*tuqa-tina	‘mother’s brother’ (Pawley 1981: 284)
Pn	Tongan	tuʔasina	‘mother’s brother’ (Douaire-Marsaudon 2015)
Pn	Tuvalu	tuātina	[MPGS etc] ‘mother’s brother’
Pn	East Uvean	tuʔasina	‘mother’s brother’
Pn	Rennellese	tuʔātina	[M{PsG}S] ‘mother’s brother’
Pn	Tikopia	tiatina	‘mother’s brother’

Blust (1980c) argues that *matuqa was the PMP term for ‘mother’s brother, wife’s father’. Blust (1994) extends the definition to include any parent-in-law. The Oceanic evidence for this extension is not substantial. Instead, MM and SES reflexes of POc *matu(q)a- more often reflect the reciprocal extension to ♂ZC. However, as many data are given a simple gloss without an expansion, one cannot put much weight on this fact.

PPn *matuqa- was the term for ‘parent’ (§2.4.1.2.3). Probably this is not derived from POc *matu(q)a ‘mother’s brother’. Rather, both terms ultimately reflect ‘old, mature (of a person)’ (vol.4:68).

PMP		*matuqa	‘mother’s brother, mother’s brother’s wife, wife’s parent’ (Blust 1980c, 1994)
PEMP		*[ma]tuqa	‘mother’s brother’ (Blust 1980c)
SH	Sawai	tua	‘mother’s brother’ (CAD)
POc		*[ma]tuqa-	‘mother’s brother’ (Milke 1958b)
MM	Usen Barok	marua-	‘mother’s brother, sister’s child’
MM	Patpatar	matuə	[FZH] ‘mother’s brother’
MM	Tolai	matua-	‘mother’s brother, ♂sister’s child’ (Fingleton 1986)
MM	Halia	tua-	‘♂mother’s brother’
MM	Hahon	coa	‘mother’s brother’
MM	Varisi	tueta-	[MFBS MFZS] ‘mother’s brother’
SES	Arosi	mau	‘mother’s brother, sister’s child’
SES	Bauro	ma-māu	‘mother’s brother, ♂sister’s child’
SES	Owa	maua-	‘mother’s brother, MMB, ♂sister’s child’
PNCV		*matu(q)a-	‘mother’s brother’ (Lynch 2001c)
NCV	Araki	nn̼ara-	‘mother’s brother’
NCV	Apma	mitue-	‘mother’s brother’
NCV	Lonwolwol	misʸo-	‘mother’s brother’
NCV	Paamese (South)	matuo-	‘mother’s brother’
NCV	Labo	mitua-	‘mother’s brother’
NCV	Rano	metuo-	‘mother’s brother’ (ACD)
NCV	Bieria	metua	‘mother’s brother’ (Lynch 2004d)
PSV		*mata-	‘mother’s brother, spouse’s father, parent’s sister’s husband’ (Lynch 2001c)
SV	Sye	meta-	[MBW BW FZH] ‘mother’s brother’
SV	Kwamera	mare-	[FZH EF P{PoG}S FFBDH GEF] ‘mother’s brother’
SV	Anejom̃	mata	[MZH EF] ‘mother’s brother’

The term *wawa occurs both as grandparent (§2.4.1.4.1) and as MB. The set supporting *wawa ‘MB’ is considerable. Possibly *wawa ‘grandparent’ and *wawa ‘MB’ were homophones with separate origins, rather than one being extended to include the other, as the sets are largely distinct in geographic distribution, and no reflex of *wawa has both senses. Terms consisting of a reduplicated syllable tend to have originated as address terms derived from the final syllable of the reference term. In this case the possible source of *wawa is phonetic rather than phonemic, as *matua was phonetically *[matuwa].

POc		*wawa	‘mother’s brother’ (Milke 1965: PNGOc)
SJ	Sobei	wawa	[MBW] ‘mother’s brother, ♂sister’s child’ (ADDR)
NNG	Gitua	wawa	‘mother’s brother’
NNG	Mangap	wo-	‘mother’s brother, o.s. sibling’s child’
NNG	Sio	wawa	‘mother’s brother, ♂sister’s son’
NNG	Avau	ava	[MBC] ‘mother’s brother’
NNG	Mengen	wowo	‘mother’s brother’
NNG	Mindiri	wau	‘mother’s brother’
NNG	Wogeo	wawa	‘mother’s brother’
NNG	Tumleo	wau, wuoyie	‘mother’s brother’ (Milke 1965: 345)
NNG	Yabem	wawa	‘mother’s brother’
NNG	Numbami	wowa	[M{PoG}S] ‘mother’s brother’
NNG	Patep	vəwa	‘mother’s brother, ♂sister’s child’
PT	Dobu	wa	‘mother’s brother’ (Milke 1965: 345)
PT	Tawala	au-, auau-	‘mother’s brother’ (aua- PL)
PT	Sinaugoro	ɣwa	[FZH FZDC] ‘mother’s brother, ♂sister’s child’
PT	Motu	vava	‘mother’s brother’ (Milke 1965: 345)
MM	Vitu	vava	‘mother’s brother, ♂sister’s child’
MM	Bola	ɣa	‘mother’s brother’
PNCV		*vʷavʷa	‘mother’s brother, father’s sister’
NCV	Raga	vʷavʷa	‘mother’s brother, father’s sister’
NCV	Big Nambas	vavei-	‘father’s sister’
NCV	Neve’ei	vavʷe	[MBW FZD] ‘father’s sister’
NCV	Nguna	wāwa	[FZH] ‘mother’s brother’ (Facey 1989)

cf. also:

NNG	Tuam	waɣa-	‘mother’s brother, (♂?)sister’s child’
PT	Iduna	ɣau-	‘mother’s brother, ♂sister’s child’
PT	Sudest	ɣa-	‘mother’s brother’

Scattered across PT and MM are reflexes of PWOc *varis. Its original meaning was presumably ‘mother’s brother’, which in accordance with cross-cousin marriage simultaneously meant ‘♂father-in-law’. Its reflexes have undergone various restrictions or extensions of meaning.

PWOc		*varis	‘mother’s brother, ♂father-in-law’
PT	Misima	valehe	‘mother’s brother’
PT	Wagawaga	warihi	‘mother’s sibling’s child’
MM	Nehan	warihi	‘♂father- or son-in-law’ (Glennon and Glennon 2006)
MM	Halia	halis	‘♂father- or son-in-law’
MM	Teop	varihi	‘wife’s brother’
MM	Hahon	waris	‘♂father- or son-in-law’
MM	Banoni	vanisi	‘parent- or child-in-law’

4.1.2.6. Father’s sister and POc *aya

Milke (1958b) believed that POc had no term for ‘father’s sister’ (FZ). Blust (1980a) argued that PMP *aya had this meaning, but this attribution is not supported by any non-Oceanic reflex glossed ‘father’s sister’. Chowning rejects PMP *aya in her comment on Blust (1980a), and omits POc *aya ‘father’s sister’ from her list of POc kinship terms in Chowning (1991). Blust and Chowning both argue their case on anthropological grounds. The case based on lexical data is presented below and shows that POc *aya seems to have been a respect term mainly denoting members of the parental generation.

Blust (1980a, 1994) has argued that PMP *aya- meant ‘FZ, FZH’. However, he admits that this is less secure than other reconstructions for members of the parental generation, and the ACD shows as broad a range of meanings. Reflexes in WMP languages have glosses that include ‘mother’s brother’, ‘father’, ‘mother’, ‘aunt’, ‘uncle’ (the last two without further specification). CMP glosses include ‘father’, ‘mother’ and ‘uncle’.³¹ Oceanic reflexes of POc *aya- fall into five groups according to their glosses: father’s sister, mother’s brother, mother, father and grandparent.

Reconstructing more than one homophonous POc term is not an option, as four of the five glosses refer to members of the parents’ generation. Blust (1980a:216) points out that if cross-cousin marriage were symmetrical and *aya meant FZ, then it would have had the secondary sense ‘spouse’s mother’, but there is no evidence that it did (cf *matuqa ‘MB’ in §2.4.1.2.5).

The ACD cites instances from geographically dispersed WMP languages and Kei (CMP) where a reflex of PMP *aya was a respect term, sometimes a term of great respect:

Tausug (NW Borneo)	aya	‘term of respect for male nobility (sultan, datuq) of the same generation as one’s father (including father)’
Iban (C Borneo)	ayaʔ	‘term of address for men of speaker’s father’s generation; term of reference for the hearer’s father’
Malay	ayah	‘father; sire; more respectful than bapaʔ’
Balinese	ayah	‘father’ (refined speech)
Dampelas (Sulawesi)	aya	‘mother’ (address form)
Makassarese (Sulawesi)	aya	‘mother’ (for persons of high rank), ‘mother’s younger sister’
Kei	yai	‘father’ (address form used by small children, and in mourning songs)

A reasonable hypothesis is that POc *aya too was a respectful address term used of a member of EGO’s parents’ generation, and that its application has tended to be narrowed in different ways in various regions. This leaves one puzzle: why do none of the reflexes embrace, say, both ‘mother’ and ‘father’, or both FZ and MB?

Several reflexes have accreted y-, but this is as expected for a term with initial *a-.

PMP		*aya	‘respect term for a member of ego’s parents’ generation’ (Blust 1980a, 1994: FZ, FZH)
PEMP		*aya-	‘father’s sister, mother’
SH	Buli	aya, aye	‘mother’ (Maan 1951)
SH	Taba	yoyo	‘father’s sister’ (Collins 1982)
Bom	Arguni	yai	‘mother’ (ACD)
CB	Ambai	ai	‘mother, mother’s sister’
CB	Wandamen	yai	‘father’
CB	Biak	yai	‘father’
CB	Serui-Laut	ai	‘mother’ (Anceaux 1961)
CB	Warembori	ai	‘father’
POc		*aya	‘member of EGO’s parents’ generation (respectful address term)’
Adm	Lele	yaya-	‘ancestor’
Adm	Titan	yaye-	[MMyZ ♀EZ] ‘mother, mother’s sister’ (Mead 1934: 218–225)
Adm	Loniu	yaya	‘mother’
NNG	Bariai	aia	‘mother, mother’s sister’
NNG	Sio	yaya	‘mother’s brother’
NNG	Sissano	aiyia	‘mother’
NNG	Roinji	yaye	‘mother’s brother’
NNG	Wab	yai	‘mother’s brother, (♂?)sister’s son’
NNG	Patep	yihi	[FZH] ‘mother’s brother’
NNG	Aria	aia	‘grandparent’
NNG	Avau	aye	[FB] ‘father’
NNG	Avau	aiyɔ	[FZC] ‘father’s sister’
PT	Dobu	yaya-	‘female of the parental generation of one’s father’s village’
PT	Ubir	ayo-	‘mother’ (anonymous Ubir wordlist)
PT	Tawala	eya-	‘father’s sister, ♂brother’s child’
PT	Sinaugoro	iaia-	[MBW] ‘father’s sister’
PT	Western Motu	lala-	[MBW ♀BC HZC] ‘father’s sister’ (Groves 1958)
MM	East Kara	yəyə	‘grandparent, grandchild’
MM	Nalik	yaya	‘grandparent, grandchild’ (Volker 1998)
MM	Siar	yai(nan)	‘mother’s grandchild’ (Rowe 2005)
MM	Babatana	zai	‘grandmother’ (McClatchey 2007)
SES	Lau	aia	‘father’s sister, ♀brother’s child’ (ADDR)
SES	Kwara’ae	ʔaiʔa	‘father’s sister’ (Deck 1934)
TM	Buma	aya	‘father’ (Tryon and Hackman 1983)
NCV	Lewo	aya	‘mother’
NCal	Xârâcùù	yāya	‘grandmother’
NCal	Kwênyii	yéyé	‘father’s sister, spouse’s mother’

Only one Oceanic FZ term is well supported, namely PPn *masakitaŋa.

PPn		*masakitaŋa	‘father’s sister’ (Marck 1996)
Proto Tongic		*mahakitaŋa	‘father’s sister’
Pn	Tongan	meheketaŋa	‘father’s sister; patrilateral female kin at G₊₁’ (Churchward 1959; Völkel 2015)
Pn	Niuean	mahakitaŋa	‘♂sister’
PNPn		*masakitaŋa	‘father’s sister’ (Marck 1996)
Pn	East Uvean	mahikitaŋa	‘father’s sister’
Pn	East Futunan	masaki[taŋa]	‘father’s sister’
Pn	Anutan	makitaŋa	‘father’s sister, father’s cousins’
Pn	Tikopia	masakitaŋa	‘father’s sister’
Pn	Pukapukan	māyakitaŋa	‘sacred maid; chief’s eldest daughter’

4.1.3. Children’s generation⇫

4.1.3.1. Child

The English term child has two distinct senses: an age-cohort sense indicating a young person’s approximate age (‘He is a child’) and a kinship sense indicating a person’s relationship to ego (‘He is my child’). The age-cohort sense is discussed in vol.5:60–65, along with the POc term *meRa ‘newborn; young person from birth to onset of adulthood’. Although reflexes of *meRa have sometimes become kinship terms, for example, in SES languages, they are not further discussed here.

POc *natu ‘child, offspring’ is widely reflected, as the cognate set below attests. It applied to ego’s children and those of s.s. siblings/parallel cousins, but not to a man’s sister’s child, on which see §2.4.1.3.2. This appears to leave a gap where a term for woman’s brother’s child is expected, but no dedicated term can be reconstructed.

POc had no dedicated terms for ‘son’ or ‘daughter’. Many Oceanic languages use the term for ‘child’ plus a term for ‘male’ or ‘female’ where specification is needed.

PEMP		*natu	‘child, offspring’ (ACD)
POc		*natu-	‘child, s.s. sibling’s child, s.s. parallel cousin’s child’ (Milke 1958b) ; ‘child’ (Ozanne-Rivierre 2000)
Adm	Mussau	natu-	[sGC] ‘child’
Adm	Baluan	naru-	[♂PBS HMBS] ‘child’ (Mead 1934: 228–342)
Adm	Titan	nat, natú-	‘child’
Adm	Pak	naro-	[♀MBC] ‘child’
SJ	Sobei	natu-	[oGC {PeG}CC P{PeG}CCC {HeB}C H{PeG}SC] ‘child’
NNG	Tuam	natu	‘child’
NNG	Sio	natu-	‘child’
NNG	Lukep	natu	‘child’
NNG	Mindiri	nalu-	‘child’
NNG	Wogeo	natu-	[sGC {PsG}CC EG EGC] ‘child’
NNG	Sissano	a-nto-	‘child’
NNG	Numbami	natu	‘child, son’
NNG	Adzera	naru-	[sGC {PsG}CC {{PoG}SC}C EoGC] ‘child’
NNG	Mangga	natu-	‘child’ (Ross’ fieldnotes)
PT	Misima	natu-	‘child’
PT	Kilivila	latu-	[♂MBC] ‘child’ (Malinowski 1929; Lounsbury 1965; Lawton f.c.)
PT	Iduna	natu-	[GC PGCC EGC] ‘child’
PT	Tawala	natu-	‘child, younger s.s. sibling’
PT	Western Motu	natu-	[{EsG}C] ‘child, s.s. sibling’s child’ (Groves 1958)
MM	Meramera	natu	‘child’ (Ross, fieldnotes)
MM	Lavongai	nat	‘child’ (Ross’ fieldnotes)
MM	Tabar	natu	‘child’
MM	Sursurunga	natu-	‘child’
MM	Siar	nat, natu-	‘child’ (Frowein 2011)
MM	Tolai	natu-	[sGC ♂ZSC ♀BSW] ‘child’ (Fingleton 1986)
MM	Banoni	nacu-	[sGC {PsG}C] ‘child’
NCV	Mota	natu-	[MBC] ‘child’ (Codrington 1891)
NCV	Raga	nitu-	[MBC] ‘child’
NCV	Araki	naru-	[BC WBC {PsG}DD FBSC WZC] ‘child’
NCV	Apma	natu-	‘child’
NCV	Daakaka	natɛ-	[MBC] ‘♂child’ (Deacon 1927; von Prince 2012: 112-117)
NCV	Paamese (South)	natu-	‘child’
NCV	Labo	nitu-	[♂BC ♂MBC] ‘child’ (Deacon 1934: 91–96)
NCV	Uripiv	natu-	[MBC] ‘child’ (Deacon 1934: 124)
NCV	Lewo	nari-	‘child’
NCV	Nguna	natu	[MBC] ‘child’ (Facey 1989)
PSV		*natu-	‘child’ (Lynch 2001c)
SV	Sye	nitu-	‘child, s.s. sibling’s child’
SV	Lenakel	nerə-	‘child’
POc		*natu-	‘child, s.s. sibling’s child, s.s. parallel cousin’s child’ (Milke 1958b) ; ‘child’ (Ozanne-Rivierre 2000)
NCal	Nyelâyu	nae-	‘child’
NCal	Cèmuhî	nai-	‘child’
NCal	Iaai	noko-	[MBC] ‘child’ (Ray 1917)
Mic	Nauruan	ŋai-	[sGC ♀BC] ‘child’ (Wedgwood 1936)
PMic		*natu	‘child’ (Hage and Marck 2002)
Mic	Kosraean	nætʌ-	‘child’
Mic	Kiribati	nāti, nati-	‘child’ (Bender et al. 2003a)
Mic	Marshallese	næci-	[♀HBC ♂BC ♂FGCC ♂MZSC ♂MBCC] ‘child’ (Spoehr 1949a)
Mic	Puluwatese	nawɨ-	[MBC] ‘child’
Mic	Satawalese	nāy, nayɨ-	[BC FGCC MBCC CC ZCC EGC] ‘child’

cf. also:

Adm	Loniu	ñɛtu	‘child’
Adm	Bipi	ñato	‘child, husband’s brother’
NNG	Roinji	nanu-	‘child’
NNG	Takia	nanu-	‘child, s.s. sibling’s child’

The Loniu and Bipi forms above, with initial ñ-, are anomalous (see Blust 1978b:48–51).

There are two variant forms of POc *natu-, both probably present in POc. The first is *qatu, the other *tu. The latter is readily understood as an abbreviation of *natú-, with stress falling on penultimate -u- when a possessor affix is added. Its Oceanic reflexes are restricted to WOc.

PEMP		*qatu	‘child, offspring’
CB	Wandamen	atu	[GC EGC] ‘child’
POc		*qatu-	‘child: offspring of ego or of s.s. sibling’
NNG	Bukawa	atu-i	‘child’ (-i PL)
MM	Bali	ɣatu	‘child’
MM	Torau	atu	‘child, s.s. sibling’s child’

PEMP		*tu	‘child, offspring’
SH	Gane	tu	‘child’ (ACD)
POc		*tu-	‘child: offspring of ego or of s.s. sibling’
NNG	Mengen	tu-	‘child’
PT	Dobu	-tu-	[{PsG}C {PoG}SC] ‘child’
MM	Vitu	tu-	‘child, sibling’s child’
MM	Bola	tu	‘child’
MM	Nehan	tu-	‘child’ (Glennon and Glennon 2006)
MM	Babatana	tu-	‘child’
MM	Marovo	tu-	[MBC] ‘child’

The generic PPn term for ‘child’ was *tama. Its reflexes in languages outside EPn indicate that more specifically it denoted ‘♀child’, and especially ‘♀son’. This implies a dedicated term for ‘♂son’ or ‘♂child’, and this space was evidently filled by PPn *fosa ‘♂son’ (Marck 1996).

PPn *tama is identical in form to the root of PPn *tama-/*tamana ‘father’ (§2.4.1.2.1). Since reciprocal uses of terms are frequent in Oceanic—‘grandparent’/‘grandchild’, ‘mother’s brother’/‘sister’s child’—it is reasonable to suppose that in some immediate ancestor of PPn the reflex of POc *tama came to be used reciprocally, and a distinction arose between suffixed *tama-/*tamana ‘father’ and unsuffixed *tama ‘child’.

PPn		*tama	‘woman or couple’s child or classificatory child, esp. son’ (Marck 1996)
Pn	Tongan	tama	[♀ZS HGC] ‘woman or couple’s child, esp. son’ (colloquially ‘fellow, lad, chap, man’)
Pn	Samoan	tama	‘♀child, ♀grandchild; child, boy; chief’
Pn	Tokelauan	tama	‘woman or couple’s child; boy’
Pn	Rennellese	tama	‘child, ♀father’
Pn	Tikopia	tama	‘child, son; sibling’s children other than irāmutu’
Pn	Rapanui	tama	‘child’ (Churchill 1912)
Pn	Māori	tama	‘son, nephew; eldest son; child (son or daughter)’
Pn	Tahitian	tama	‘child, young people in general’
Pn	Hawaiian	kama	‘child, person’

From the kinship term *tama evidently developed the PPn age cohort terms *tama-qiti ‘child’ and *tama-riki ‘children’ (Marck 1996), the second element of each was the reflex of a POc term for ‘small’ (vol.2:193–194). Additionally PPn innovated the term *fosa ‘♂son’ (Marck 1996).

PPn		*fosa	‘♂son, ♂brother’s son’ (Marck 1996)
Pn	Tongan	foha	[♂BS ♂WGS] ‘♂son’ (Marck 1996; Völkel 2015)
Pn	East Uvean	foha	[BS] ‘♂son’ (Tongan loan?) (Burrows 1938)
Pn	East Futunan	vosā	[GS] ‘son’ (Burrows 1936)
Pn	Rennellese	hosa	[♂BS ♂MBSS ♂F{PsG}CSS ♂{PoG}DS ♂{PsG}CDS] ‘son’
Pn	Tikopia	fosa	[♂BC] ‘♂child’ (some speakers hold that fosa applied only to brother__s child)

4.1.3.2. ♂Sister’s child

POc *[qa]lawa has been regularly offered in the literature as the term for ‘♂sister’s child’ (♂ZC). A number of its reflexes are glossed ‘mother’s brother’ (MB), whilst three widely spaced languages—Wogeo, Nakanai and Nauruan—use it reciprocally for both ♂ZC and MB (the Nakanai form is under ‘cf. also’ because it is not a regular reflex). If extended senses were recorded for all reflexes, one could be more confident about the gloss of the POc reconstruction. As things stand, POc *[qa]lawa probably meant ♂ZC, but because ♂ZC/MB was a recognised relationship from very early Oceanic, one cannot be sure whether it already also referred to MB in POc.

POc		*[qa]lawa	‘♂sister’s child, (?) mother’s brother’ (Pawley 1981:284: *(qa)lawa; Milke 1958b: *(a)lava; 1968: *qalawa)
Adm	Mussau	aloa-	[FZH oGC] ‘mother’s brother’
SJ	Sobei	ewo-	[MBW ♂PGDC ♂PPGCD ♀HZC ♀HPGDC] ‘mother’s brother, ♂sister’s child’
NNG	Wogeo	kalawa	[♂FBDC ♂MZDC ♂MBDC ♂FZDC] ‘mother’s brother, ♂sister’s child’
NNG	Manam	elua	[MBS etc] ‘mother’s brother’
PT	Sudest	loɣai	‘mother’s brother’ (Lepowsky 1981)
MM	Meramera	loa	‘mother’s brother’ (Ross, fieldnotes)
MM	Patpatar	lauə	‘♂sister’s child’
MM	Ririo	aluwa	‘♂sister’s child’ (Milke 1958b)
PNCV		*(q)alawa	‘sister’s child’ (Clark 2009: *alawa)
NCV	Lombaha	aloa-	‘♂sister’s child’
NCV	Raga	aloa	‘♂sister’s child’
NCV	Araki	elua-	[FZW HMB FBDS MZDS] ‘mother’s brother’
NCV	Tamambo	alua	‘sister’s child’
NCV	Apma	oloa-	‘sister’s son’
NCV	Labo	lowo-	‘♂sister’s child’
NCV	Lendamboi	alawa	‘sister’s child’
NCV	Nguna	aloa	[FZH MZH] ‘mother’s brother’ (Guiart 1964)
PSV		*alwə-	‘♂sister’s son’ (?) (Lynch 2001c)
SV	Sye	alwo-	‘cross-cousin’ (Lynch 2001c)
SV	Ura	alwi-	‘nephew’
Mic	Nauruan	aroe-	‘mother’s brother, ♂sister’s child’ (Wedgwood 1936)

cf. also:

MM

Nakanai

hala-

‘mother’s brother, ♂sister’s child’ (ADDR; for †halaua-)

Micronesian and Central Pacific languages share a term reflecting a putative POc *pa(s,c)u/*pa(s,c)ua- ♂sister’s child—“putative” because there is only one reflex outside Mic and CP languages, and because there is some evidence that the term refers or referred to the relationship between MB and ♂ZC rather than to the person of the ♂ZC (Douaire-Marsaudon 2015; Pauwels 2015). The items listed under ‘cf. also’ are not cognate with the set above them, but are perhaps the result of borrowing.

POc		*pa(s,c)u, *pa(s,c)ua-	‘♂sister’s child’ (?)
MM	Hahon	pasu-	‘sister’s child’
PMic		*fa(s,S)ua	‘o.s. sibling’s child’ (Bender et al. 2003a)
Mic	Satawalese	fatɨw, fatɨwa-	[MZDC MMZDDC] ‘sister’s child’
Mic	Woleaian	fatuwe-	‘sister’s child’ (Burrows and Spiro 1957)
Mic	Ulithian	faθie	‘sister’s child’
PCP		*vasu	‘♂sister’s child’ (Hage and Marck 2002)
Fij	Tavuki	vasu	‘sister’s child’
Fij	Bauan	vasu	‘sister’s child’ (Capell 1941)
PPn		*fasu	‘♂sister’s child’ (Hage and Marck 2002)
Pn	Tongan	fahu	‘sister’s child’ (Douaire-Marsaudon 2015)

cf. also:

Fij	Wayan	batuvu	‘sister’s child’
Fij	Tavuki	vatuvu	‘sister’s child’
Fij	Tokatoka	vatuvu	‘sister’s child’

Another regional set consists of reflexes of Proto Central Vanuatu *matailau ‘♂sister’s child’. The term appears to break down into *mata i lau ‘eye of the sea’, but this makes no obvious sense.

Proto Central Vanuatu		*matailau	‘♂sister’s child’
NCV	Apma	metulu	‘sister’s son’
NCV	Daakaka	mɛtɔ̄	‘sister’s child’ (Deacon 1927; von Prince 2012: 112-117)
NCV	Lonwolwol	metelo	‘♂sister’s son’
NCV	Paamese (South)	meteilau	‘sister’s son’
NCV	Big Nambas	məriri-	‘sister’s child’
NCV	Neve’ei	metelau	[WBC] ‘sister’s child’

The PPn term for ‘♂sister’s child’ was *qilāmutu. It has no known non-Pn reflexes, but is widely reflected in Polynesia. The Tuvalu, Takuu, Nukumanu and Luangiua reflexes also mean ‘mother’s brother’.

PPn		*qilāmutu	‘♂sister’s child’ (Marck 1996: *ʔilamutu)
Pn	Tongan	ʔilamutu	‘♂sister’s child’ (Völkel 2015)
PNPn		*qilāmutu	‘♂sister’s child’ (Marck 1996)
Pn	Samoan	ilāmutu	‘descendants in female line’
Pn	Tokelauan	ilāmutu	‘♂sister’s child; ♂descendants of sister’s children; any kinsman who is an offspring of a female sibling of any male forebear’ (Huntsman 1971)
Pn	East Futunan	ilamutu	‘♂sister’s child’ (Burrows 1936)
Pn	Rennellese	ʔiŋāmutu	‘♂sister’s child, ♂father’s parallel cousin’s daughter’s child’
Pn	Tikopia	iramutu	‘♂sister’s child’
Pn	Pileni	ilamutu	‘♂sister’s child’
Pn	Takuu	[i]lāmotu	‘mother’s brother; ♂sister’s child’
Pn	Luangiua	lamoku	‘mother’s brother; ♂sister’s child’
PEPn		*irāmutu	‘sibling’s child’ (Marck 1996)
Pn	Māori	irāmutu	‘sister’s child (traditional sense)’
Pn	Marquesan	iʔamutu	‘sibling’s child’
Pn	Mangarevan	iramutu	‘sibling’s child’ (Hīroa 1938:136)

4.1.4. Grandparent and grandchild generations and beyond⇫

4.1.4.1. Grandparent, grandchild (reciprocal)

The terms reconstructed in this subsection were evidently used reciprocally in POc, as they are in many modern Oceanic languages. Many of them are glossed ‘grandparent, grandchild’ (PP, CC), but others have a narrower definition, e.g. ‘mother’s father’. Here one is again up against the fact that extended senses are often not recorded. It is likely that many more should have been recorded as ‘grandparent, grandchild’. Many ethnographers indicate that these terms apply to all (or almost all) members of the grandparent and grandchild generations, so that a wider definition, ‘kin of the grandparent and grandchild generations’, would in many cases be appropriate. In some languages the definition is wider still, as the term applies to ancestors including grandparents and all their forebears (see §2.4.1.4.3).

Members of the grandparent and grandchild generations are not usually differentiated by gender, but sometimes a modifier is added. This seems to be particularly the case in EPn languages, e.g., Hawaiian kupuna kane ‘grandfather, grand-uncle’ vs kupuna wahine ‘grandmother, grand-aunt’, and cognate forms in other EPn languages.

In some languages a grandparent term is used for the mother’s brother, as the glosses indicate. Blust (1980a:214–215) sees the association as ‘due to the notion of kin-based sovereignty common to the life-giving clan ancestors (*ampu, etc.) and the life-perpetuating wife-giving group (*ma[n]tuqa), i.e. a matter of power’.

Terms apparently reflecting POc †*tibu- are also included here (see discussion in §2.4.1.1).

PMP		*tumpu	‘grandparent, grandchild’ (ACD)
POc		*tubu-	‘grandparent, grandchild’ (?) (ACD; Milke 1958b: ‘grandparent’)
Adm	Mussau	tuvu-	‘grandparent, grandchild’
Adm	Wuvulu	ʔupu-	‘grandparent, grandchild’
Adm	Baluan	tupu-	‘all relatives in the grandparent generation’
Adm	Lele	tuʙu	[MF MMB] ‘mother’s brother’’ (Mead 1934: 345–347)
Adm	Nyindrou	tubu-	‘grandparent’
SJ	Sobei	tapu-	‘((great-)great-)grandparent’
NNG	Malalamai	tiᵐbu	‘mother’s brother’
NNG	Bariai	tibu-, i-tub	‘grandparent, grandchild’ (i-tub 3SG)
NNG	Mangap	tumbu-	‘grandparent, grandchild’
NNG	Sio	timbu	‘grandparent, grandchild’
NNG	Lukep	tibu-	‘grandchild’
NNG	Atui	tivi-	‘grandparent’
NNG	Mengen	sivu-	‘grandchild’
NNG	Takia	tbu-	‘grandparent, grandchild’
NNG	Manam	tubu-	‘grandparent, grandchild’
NNG	Yabem	dɪbu-	[CS MMZH] ‘parent’s father’
NNG	Yabem	dɪbu-ò	[CD FFBW] ‘parent’s mother’ (-ò FEM)
NNG	Wampar	roᵐpo-	‘grandparent, grandchild’
NNG	Numbami	tubu	‘grandparent’
PT	Sudest	rumbu-	‘grandparent, grandchild’
PT	Dobu	tubu-	‘grandparent, grandchild’
PT	Sinaugoro	tubu-	‘grandparent, grandparent’s sibling, grandchild’
PT	Roro	kupu-	‘grandparent, grandchild’
MM	Vitu	tubu	[PPP CCC] ‘grandparent, grandchild’
MM	Bola	tubu	‘grandparent, grandchild’
MM	Lavongai	tivu	‘grandparent, grandchild’ (Fast and Fast 1989)
MM	Siar	tubu-	‘grandparent, grandchild’ (Frowein 2011)
MM	Tolai	tubu-	[FZ FZD FZDD MBDS] ‘grandparent, grandchild’ (Fingleton 1986)
MM	Nehan	tubu-	[PPP CCC] ‘grandparent’ (Glennon and Glennon 2006; Nachman 1978)
MM	Halia	tubu-	‘grandparent, ancestor’
SES	Bugotu	tubu-	‘mother’s brother, ♂sister’s child’ (Ivens 1940)
SES	Lengo	tubu-	‘mother’s brother, ♂sister’s child’
SES	Bauro	wa-upu	‘mother’s brother’ (wa-‘male’)
PNCV		*tubu-	‘grandparent, grandchild’ (Clark 2009)
NCV	Mota	tupu-	[MBCC FZCC] ‘grandparent, grandchild’ (Codrington 1891)
NCV	Lolovoli	tubu-	‘grandfather’
NCV	Raga	sibi	‘mother’s father, sister’s husband’
NCV	Araki	tapu-	‘grandfather’
NCV	Sa	tibi-	‘grandfather’
NCV	North Ambrym	tuvyu-	[PPE CE] ‘grandparent, grandchild’ (Franjieh 2012:226, 371; Löffler 1960)
NCV	Paamese (South)	tevi-	‘grandparent’
NCV	Neve’ei	cube-	‘grandparent, grandchild’
NCV	Lewo	repi-	‘grandparent’
NCV	Nguna	topu-	‘grandfather’
PSV		*[e]t(p,b)u-	‘grandparent’ (Lynch 2001c)
SV	Sye	re-tpo-	‘wife’
SV	Lenakel	rəpə-	‘grandparent’
SV	Anejom̃	e-tpo-	‘grandparent’
PNCal		*tuᵐbu-	‘grandparent’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	kibu-	‘grandparent’
NCal	Fwâi	sıũ̃-	‘grandparent’
NCal	Kwênyii	tũ̃-	‘grandparent, mother’s father, spouse’s father’
NCal	Iaai	kibe-	‘grandparent’
PMic		*tupʷu	‘grandparent, grandchild’ (Hage and Marck 2002)
Mic	Kiribati	tipu-	‘grandparent, grandchild’ (Groves et al. 1985)
Mic	Marshallese	cipʷi-	‘grandparent, grandchild’ (Hage and Marck 2002)
Mic	Woleaian	subu	‘grandparent, grandchild’ (Hage and Marck 2002)
Fij	Wayan	tubu-	‘grandparent, classificatory grandparent, ancestor’
Fij	Tokatoka	tubu-	[FFB FFFBS FFF etc] ‘father’s father’
PPn		*tupu-, *tupuna	‘kin of the second and further ascending generations’ (Marck 1996)
Pn	Niuean	[matua] tupuna	‘ancestor’ (matua ‘parent’; Sperlich 1997)
PNPn		*tupu-, *tupuna	‘kin of the second and further ascending generations’ (Marck 1996)
Pn	Tuvalu	tupuna	‘grandparent, grand-uncle/-aunt’
Pn	Rennellese	tupuna	‘(great-)grandparent’
Pn	Pileni	tʰupu-	‘grandparent (ADDR)’
Pn	Takuu	tipuna, tippuna	‘relative by blood or marriage two or more ascending generations’ (tippuna PL)
Pn	Pukapukan	tupuna, tūpuna	‘ancestor, grandparent’ (tūpuna PL)
PEPn		*tupuna	‘kin of two or more ascending generations’ (Marck 1996)
Pn	Rapanui	tupuna	‘grandparent’ (Churchill 1912)
Pn	Māori	tupuna, tipuna	‘grandparent, ancestor’
Pn	Tahitian	tupuna	‘grandparent, ancestor’
Pn	Marquesan	tupuna	‘(great-)grandparent, etc’
Pn	Hawaiian	kupuna	‘grandparent’

It was noted in §2.4.1.1 that there are also the variants *ubu- and *kubu-. The latter has only two reflexes, both kubu- ‘grandparent, grandchild’, in two closely related PT languages, Bwaidoga and Iduna. The former is a little more convincingly attested.

PMP		*umpu	‘grandparent, grandchild’ (ACD)
POc		*ubu	‘grandparent, grandchild’
MM	Torau	up-	‘grandparent’
NCV	Neve’ei	vubʷu	[FFZS] ‘grandfather’
Mic	Mortlockese	upʷu(-tiw)	‘grandparent, grandchild’ (Hage and Marck 2002)

In §2.4.1 forms were noted form reflecting POc *tabu- and *abu. They do not differ in meaning from *tubu- and *ubu, and no known phonological rule accounts for them. The ACD records no reconstruction earlier than PEMP.

PEMP		*tampu	‘grandparent, grandchild’ (ACD)
POc		*tabu-	‘grandparent, grandchild’
Adm	Ponam	tāʙu	‘grandparent, ancestor’
Adm	Leipon	jābu	‘grandparent, ancestor’
NNG	Tuam	tapu-	[PGS PPGCS] ‘father, parent’s brother’
NNG	Sissano (Arop)	tapu-	‘grandparent’
MM	Tangga	tabu- -lik	‘grandchild’
MM	Tangga	tabu- -tamat	‘mother’s father’
NCV	Araki	tapu-	‘grandparent, grandchild’
NCV	Neve’ei	tabu-	[FFZS] ‘grandfather’

PMP		*ampu	‘grandparent, grandchild’ (ACD)
POc		*abu[a]	‘grandparent, grandchild’ (ACD)
Adm	Baluan	apua	‘members of G₊₃ and G₋₃’
NNG	Tuam	abu	‘grandparent, grandchild’
NNG	Mangap	abu	‘grandmother, granddaughter’
NNG	Sio	ābu	‘grandchild’
NNG	Amara	avu	‘mother’
NNG	Kis	abu	‘elder o.s. sibling’ (Z’graggen 1974b)
NNG	Bukawa	abu-ŋgaʔ	‘grandfather, grandson’
NNG	Bukawa	abu-wɪ	‘grandmother, granddaughter’
NNG	Numbami	abu	‘grandparent’
PNCV		*abu[a]	‘grandparent, grandchild’
NCV	Apma	apu	‘grandparent’
NCV	Paamese (South)	avue	‘grandparent, spouse’s parent’
NCV	Uripiv	apu	‘grandparent’
NCV	Lewo	apua	‘grandparent’
NCV	Nguna	pua	[MFB] ‘mother’s father’ (Facey 1989)
NCal	Iaai	ɔ̄bʷi-	‘grandchild’
NCal	Nengone	abu-n	‘grandchild’ (Ozanne-Rivierre 2000)

The ancestors of POc *bubu date back to PAn. Historically they have nothing to do with the forms above, but *bubu now fits into the quasi-paradigm of terms for blood relatives described in §2.4.1.1 and in some languages serves as an address term.

Reflexes in languages that have lost final *-u have become *[ᵐbuᵐb] and then lost the final *[b] giving the forms in Wogeo, Ulau-Suain, Sissano and Avava. Other languages reflect an abbreviation of *bubu to *bu.

PAn		*bubu	‘grandparent, grandchild’ (ACD)
PMP		*bubu	‘grandparent, grandchild’ (ACD)
POc		*bubu	‘grandparent, grandchild’ (Blust 1980a; Chowning 1991)
NNG	Sio	mbupu	‘grandchild’
NNG	Mindiri	bu-	‘great-grandparent’
NNG	Wogeo	bum	‘grandfather’
NNG	Sissano	(a)pum	‘grandfather’
NNG	Adzera	bu-	‘spouse’s parent, child’s spouse’
NNG	Mapos Buang	bu	‘grandparent, grandchild’
PT	Sudest	bubu	‘grandparent, father’s brother’
PT	Kilivila	bubu	‘grandparent, grandchild’ (Lawton f.c.)
PT	Molima	bubu	‘grandparent’
MM	East Kara	bu-	‘ancestor’
MM	Petats	bubu	‘grandparent, grandchild’
NCV	Mota	pupu-a	‘grandparent, grandchild’
NCV	Lolovoli	bubu	‘grandparent’
NCV	Raga	bibi	[PFZ H] ‘grandfather’
NCV	Apma	bua-	‘grandparent’
NCV	Neve’ei	bubu	‘grandparent, grandchild’
NCal	Iaai	buba-	‘grandparent’
Mic	Marshallese	pɨpɨ	[PGCC] ‘child’s term for grandfather’ (Spoehr 1949a)
Fij	Bauan	bu-	‘grandfather’
Pn	West Futunan	pu-a	‘grandparent’
Pn	Pileni	pu-	‘grandparent, grandparent’s sibling, ancestor’

As noted in §2.4.1.2.2 with regard to *inai ‘mother (ADDR)’, Blust (1979) reconstructs a set of PMP address terms in *-y, continued as POc *-i. POc *[bu]bui is also one of these, and it too fails to undergo a regular sound change. Normally, PMP *-uy becomes POc *-i (§1.3.4.2), but this seems to be another instance of an address term resisting change (cf. §2.4.1.2.2).

The reflexes under ‘cf. also’ refer to female kin of the parent generation, and have presumably come to be applied to them as a matter of respect. The ra- of Merei ra-ᵐbui- is a respect prefix used on terms for senior women (§2.4.1.2.2).

PMP		*bubu-y	‘grandparent, grandchild’ (Blust 1979)
POc		*[bu]bui	‘grandparent, grandchild’
NNG	Takia	bui	‘grandchild’
MM	Lavongai	vuvui	‘grandparent, grandchild’ (Fast and Fast 1989)
NCV	Nokuku	pui-	‘grandparent, grandchild’
Fij	Moala	bui	‘grandmother’; ‘wife’ (rarely)

cf. also:

NCV	Longana	bui	‘mother, mother’s sister’
NCV	Wusi	pui	‘mother’(Lynch 1996)
NCV	Merei	ra-ᵐbui-	‘mother’(Lynch 1996)
NCV	Akei	bʷi-	‘mother’s brother’

POc *wawa[-] ‘grandparent, grandchild’ is of only PEMP antiquity, does not belong to the paradigm in §2.4.1 and is homophonous with POc *wawa ‘mother’s brother’ (§2.4.1.2.5). The reduplicated form suggests that it probably originated as an address term.

PEMP		*wawa	‘grandparent, grandchild’
CB	Ambai	wawa	‘ancestor’
POc		*wawa	‘grandparent, grandchild’
NNG	Singorakai	wau	‘grandparent, grandchild’
PT	Ubir	wawa-	‘grandfather’ (Ubir wordlist)
PT	Dawawa	waɣa	‘(great-)great-grandparent’
MM	Sursurunga	wowo	‘mother’s mother, o.s. sibling’
MM	Siar	wɔwɔ	‘grandmother’ (Frowein 2011)
MM	Varisi	vavae	‘grandfather, male of G₊₂’
SES	Longgu	vua	‘grandparent, grandchild’
SES	’Are’are	wauwa-	‘grandparent, grandchild’
SES	Sa’a	waue[-]	‘grandfather, grandson’
SES	Bauro	wa-uwa	[ESG PGESC] ‘grandfather’ (wa- MALE)
SES	Owa	wauwa	‘grandfather’
SES	Owa	wauwa-	‘grandson’
NCal	Dehu	ww̥aww̥a	‘grandfather’ (Ray 1917)

4.1.4.2. Grandchild (non-reciprocal)

Although a grandchild is often referred to by the reciprocal grandparent/grandchild terms listed in the previous subsection one non-reciprocal term is found, POc *makubu- ‘grandchild’, or more properly ‘kin two generations below ego’. It is not reflected in WOc or SES, but there are reflexes in Admiralties languages, as well as non-Oceanic cognates, listed in the notes beneath the ACD’s entry for PMP *empu. These are from three Sulawesi languages: Bare’e makumpu opu ‘great-grandchild’, Dampelas maʔupu ‘grandchild’ and Uma kumpu ‘grandchild’, this last described as a ‘foreshortened form’.

The ACD’s reconstruction is POc *mo-kobu, ‘with regressive assimilation of the remaining first syllable vowel’, based on the Pak and Pn forms cited below. There is no doubt that assimilation occurred, but later and separately in the Admiralties and PPn. The cognate set below supports POc *ma-kubu- fairly strongly. This form evidently includes as its stem POc *kubu- (Table 2.12), reflected unprefixed only in the Bwaidoga and Iduna reflexes noted in the previous subsection.³²

POc		*makubu-	‘grandchild; kin two generations below ego’ (ACD: *mo-kobu)
Adm	Pak	mo-kopu-	‘grandchild’
Adm	Lele	meŋʙu	‘grandchild’
PNCV		*makubu	‘grandchild’ (Clark 2009)
NCV	Apma	meʙu-	‘grandchild’
NCV	North Ambrym	mabœɔ-	‘great-grandson’ (Guiart 1951)
NCV	Big Nambas	nn̼əxitt̼-	‘grandchild’
NCV	Labo	neiʙü-	[MBDS ♂MBSS] ‘grandchild’
NCV	Unua	meɣeʙu-	‘grandchild’
NCV	Port Sandwich	mexiᵐbü-	‘grandchild, descendant’
PSV		*maɣubʷu-	‘grandchild’ (Lynch 2001c)
SV	Sye	moɣpo-	‘grandchild’
SV	Lenakel	mʷipʷə-	‘grandchild’
SV	Southwest Tanna	makupu	‘grandchild’
Fij	Tavuki	makubu	[BCC MBCC] ‘grandchild’
Fij	Bauan	makubu-	‘grandchild’
PPn		*makupu-na	‘kin of two or more descending generations’ (Marck 1996)
Pn	Tongan	mokopu-na	‘grandchild’
Pn	Rennellese	makupuna	‘grand-child/-niece/-nephew’
Pn	Pileni	mokupuna	‘grandchild; member of grandchild generation and lower’
Pn	Pukapukan	makopuna	‘grandchild’
Pn	Rapanui	makupuna	‘grandchild’ (Churchill 1912)
Pn	Māori	mokopuna	‘grand-child/-niece/-nephew’
Pn	Marquesan	moupuna	‘grand-child/-niece/-nephew’
Pn	Hawaiian	moʔopuna	‘grandchild’

cf. also:

PT	Bwaidoga	kubu-	‘grandparent, grandchild’
PT	Iduna	kubu-	‘great-grandparent, great-grandchild’

4.1.4.3. Generations beyond grandparent and grandchild

Just one POc term is reconstructable for an ascending generation beyond the grandparent generation and its reciprocal. This is POc *bawa[-], for which reflexes have been found in only a few languages.

PEMP		*bawa	‘great-great-grandparent, great-great-grandchild’
RA	Ambel	baw	‘member of G₊₄ or G₋₄’
POc		*bawa[-]	‘great-great-grandparent, great-great-grandchild’
Adm	Sori-Harengan	(ña)bʷau	[FFB MZH] ‘grandfather’ (ña- is prefixed to some kin terms)
Adm	Pak	pʷawa(riu)	‘great-grandfather’
PT	Sinaugoro	bava-	‘member of G₊₅’
MM	Nakanai	bao	‘member of G₊₄ or G₋₄’

There are probably a number of reasons why terms for these relations are not found. One is that in many Oceanic languages reflexes of POc *tubu- (§2.4.1.4.1) also mean ‘ancestors, descendants’, i.e. not only grandparents and grandchildren but also great-grandparents and great-grandchildren and the generations beyond.

If there were once distinct lexical items for generations beyond PPP and CCC, they have been replaced in various Eastern Oceanic languages by a phrase that allows the precise generation to be stated, as seen in these examples:

SES	Sa’a	waue haʔa-rua	‘great-grandfather’
SES	Sa’a	waue haʔa-olu	‘great-great-grandfather’
NCV	Araki	tapu-ku ð̼aha-dua	‘my great-grandfather’
Fij	Wayan	tubu vaka-tolu	‘member of G₊₅’
Fij	Nadrogā	tai vā-rua	‘great-grandfather, great-grandchild’

The structure consists of the term for ‘grandparent, grandchild’, followed by a multiplicative numeral formed with a prefix reflecting PEOc *paka- ‘causative’ (§14.5.2). As François (2002) notes of the Araki form, ‘my great-grandfather’ is labelled ‘my grandfather twice’. A similar strategy occurs in Pn languages, but without the causative morpheme: thus Tongan kui-ua, Manihiki-Rakahanga tupuna tua-rua, Hawaiian kupuna kua-lua, all ‘grandparent [xxx]-two, ‘great-grandparent’, and correspondingly kui-tolu, tupuna tua-teru, kupuna kua-kolu ‘grandparent [xxx]-three, ‘great-great-grandparent’ (Marck 1996:211).

In languages close to the New Guinea mainland, however, it seems that there were terms for distinct generations, but, other than POc *sese- none can be reconstructed beyond a very local level.

POc		*sese-	‘great-great-grandparent (?)’
Adm	Lele	sese-	‘grandparent’
NNG	Kove	sese-	‘member of G₊₄’ (Chowning 2009)
PT	Sinaugoro	sese-	‘member of G₊₄’

PNNG		*sasa	‘great-grandparent, great-grandchild’
NNG	Tuam	sās	‘member of G₊₄ or G₋₄’
NNG	Bariai	sasa	‘member of G₊₄’
NNG	Mangap	sāza	‘great-grandparent, great-grandchild’
NNG	Amara	asasa	‘great-grandparent, great-grandchild’
NNG	Lukep	sassa	‘great-grandparent’
NNG	Aria	sasa	‘member of G₊₄’

As noted in §2.4.1.5.1, a great-grandparent is in scattered languages denoted by the term *taci- ‘younger s.s. sibling’.

4.1.5. Ego’s generation: siblings and cousins⇫

4.1.5.1. Same-sex sibling, younger same-sex sibling

POc *taci- ‘younger s.s. sibling’ is reflected throughout Oceanic. It forms a pair with POc *tua-, *tuaka ‘elder s.s. sibling’ (§2.4.1.5.2), but *taci is what Greenberg (1980) and Hage (2001b) would call the ‘unmarked’ member of the pair. That is, if a term is needed to denote the category ‘s.s. sibling’, *taci- will be chosen. This also means diachronically that a reflex of *taci- sometimes comes to denote ‘s.s. sibling’, but a reflex of *tua-, *tuaka never does.

Occasionally a kinship terminology loses the distinction between s.s. sibling and o.s. sibling but retains the distinction between younger and elder. In this case, the terms for ‘s.s. sibling’ are unmarked, and a reflex of *taci- denotes all younger siblings and a reflex of *tua-, *tuaka all elder siblings.³³

Because of bifurcate merging (§2.3.1) in many languages a reflex of POc *taci- is also used of a s.s. parallel cousin. In some of these languages, e.g. Yabem, the reflex means that it is the cousin who is younger than EGO. In other languages, e.g. Tongan, it indicates that the cousin’s parent is younger than EGO’s parent.

Another extended meaning is ‘s.s. sibling’s spouse’, or sometimes ‘younger s.s. sibling’s spouse’, discussed in §2.4.2.2.

PMP		*taji	‘younger s.s. sibling’ (ACD; see also §2.4.2.2)
POc		*taci-	‘younger s.s. sibling, parallel cousin’s younger s.s. child’ (Milke 1958b)
Yap	Yapese	teθī-	‘younger brother’ (Jensen 1977)
Adm	Mussau	tasi-	[sGE] ‘wife’s sister, husband’s brother’
Adm	Pak	tehi-	[{PsG}sC ♂{PoG}SS] ‘s.s. sibling’
Adm	Lele	dere	‘s.s. sibling’
SJ	Sobei	tasi-	‘younger sibling’
NNG	Bariai	tadi-	[PGsC] ‘s.s. sibling’
NNG	Amara	tei-	[{PsG}sC] ‘s.s. sibling’
NNG	Atui	tei-	‘s.s. sibling’
NNG	Mengen	taitai	‘younger s.s. sibling’
NNG	Wab	tai	‘younger s.s. sibling’ (Z’graggen 1974a)
NNG	Takia	tei-	[EoG sGE] ‘s.s. sibling’
NNG	Manam	tari	‘younger sibling’
NNG	Kairiru	tei-	‘sibling, wife’s sibling’
NNG	Yabem	lasi-, lasi-ò	[{PsG}ysC] ‘younger s.s. sibling’ (-ò FEM)
NNG	Adzera	rasi-	‘grandparent, grandchild’
NNG	Mapos Buang	ari-	‘s.s. sibling’
PT	Misima	tali-	‘s.s. sibling’
PT	Dobu	tasi-	[{PsG}sC GsE {PsG}sCE {PoG}sCE] ‘s.s. sibling’
PT	Sinaugoro	tari-	‘younger sibling’
PT	Western Motu	tadi-	‘younger sibling’ (Groves 1958) ; ‘younger s.s. sibling’ (Lister-Turner & Clark 1954)
MM	Vitu	taði-	‘s.s. sibling’
MM	Bola	tari-	‘younger s.s. sibling’
MM	Lavongai	təsi	‘brother’ (Fast and Fast 1989)
MM	Nalik	dasi-	‘(?♂)brother’ (Chinnery 1929)
MM	Siar	tasi-	‘sibling, cousin’ (Frowein 2011)
MM	Patpatar	təsi-	‘brother, mother’s sister’s son’
MM	Roviana	tasi-	‘s.s. sibling’ (Capell 1943b)
MM	Maringe	tahi-	[{PsG}yC] ‘younger s.s. sibling’
SES	Bugotu	tahi-	‘s.s. sibling (younger?)’ (Ivens 1940)
SES	To’aba’ita	θa-asi-	‘younger sibling’
SES	Kwaio	asi-	‘younger brother, younger male cousin’
SES	Bauro	k-asi, w-asi	‘younger s.s. sibling’ (k-female, w-male)
NCV	Mota	tasi-	[MZSC FBSC] ‘younger s.s. sibling’ (Codrington and Palmer 1896)
NCV	Longana	tehi-	[FFC MMC] ‘younger sibling’
NCV	Araki	rasi-	[WZyH HyB] ‘younger brother’
NCV	Apma	tasi-	‘younger sibling’
NCV	Paamese (South)	tasi-	‘younger sibling’
NCV	Neve’ei	tasu-	[{FyB}sC] ‘younger s.s. sibling’
NCV	Uripiv	tasi-	‘younger sibling’
NCV	Nguna	tai	[{PsG}S] ‘s.s. sibling’ (Loss of *-s- is unexpected)
SV	Lenakel	no-rhə-	‘younger s.s. sibling’
PNCal		*tasi-	‘younger s.s. sibling’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	kāri-	‘younger s.s. sibling’
NCal	Drubea	tī-	‘younger s.s. sibling’
NCal	Iaai	kei-	‘younger s.s. sibling’
Mic	Kiribati	tari	[{PsG}sC] ‘s.s. sibling’ (Lambert 1981)
Fij	Wayan	taði-	[{PysG}C] ‘younger s.s. sibling, sibling’
Fij	Deuba	taði-	[{PysG}C CCC] ‘younger s.s. sibling’
PPn		*tahi-, *tahina	‘younger s.s. sibling’ (Marck 1996)
Pn	Tongan	tehina	[{PysG}sC] ‘younger s.s. sibling’ (Völkel 2015)
Pn	Niuean	tehina	‘younger s.s. sibling, parallel cousin’
PNPn		*tai-, *taina	‘younger s.s. sibling’ (Marck 1996)
Pn	Samoan	tei	‘younger s.s. sibling, s.s. cousin’ (Williamson 1924)
Pn	Tuvalu	taina	[EoG EoGE] ‘s.s. sibling, s.s. cousin’
Pn	Rennellese	taina	[MBysC] ‘younger s.s. sibling’
Pn	Rapanui	teina	‘younger sibling, younger cousin’ (Churchill 1912)
Pn	Tahitian	teina	[{PyG}C] ‘younger s.s. sibling’
Pn	Marquesan	teina	‘younger s.s. sibling’
Pn	Hawaiian	kai-kaina	‘younger sibling, younger cousin’

Two further variants are reconstructable, namely POc *aci and *kaci- ( §2.4.1.1). The former has non-Oceanic reflexes (ACD) but only one known Oceanic reflex, Torau asi- ‘younger s.s. sibling’. POc *kaci- has the reflexes listed below.

POc		*kaci-	‘younger s.s. sibling’
PT	Bwaidoga	kai-	[PyGC] ‘younger sibling’
MM	Banoni	kasi-	‘younger s.s. sibling’
MM	Uruava	kái-	‘younger s.s. sibling’
MM	Babatana	kəe-	‘younger s.s. sibling’
SES	Birao	kasi-	‘younger s.s. sibling’

4.1.5.2. Elder same-sex sibling

As indicated in the overview in §2.3.5, POc had three terms for ‘elder s.s. sibling’. POc *tua- was a reference term that took a possessor suffix, and *tuaka, perhaps originally an address term, is also reflected as a reference term in many Oceanic languages. Both were POc innovations, as no convincing non-Oceanic cognates have been found. The third term, *kaka, was—at least originally—an address term, of PAn antiquity.

It follows from bifurcate merging that the terms below also referred to an elder parallel cousin, but there are also languages (Roro, Neve’ei, Wayan Fijian, Tahitian) in which the ‘elder’ feature belongs not to the cousin but to the cousin’s uncle, i.e. the extended meaning of the term is not {PsG}eC but {PseG}C, ‘father’s elder brother’s child’.

Some previous writers have reconstructed *tuaka as *tuqaka (§2.4.1.1). No reflex in our data supports *-q-. No reflex attesting to both *-q- and *-k- has been found. Instead, the data reflect *tua- or *tuaka-, listed separately below. Many reflexes of *tuaka- show coalescence of *-ua- to -o-, giving toka-, and a few show deletion of *-a-, giving tuka-, both probably a result of stress shifting to the right when a possessor suffix was added, e.g. *tuaká-gu ‘my …’, perhaps aided by the tendency for POc roots to be disyllabic.

Why are there two reconstructable forms? A number of Oceanic subgroups have both forms, suggesting that there is/was a functional distinction between them. Several languages have reflexes of both forms, the distinction between them varying from language to language, making it impossible to reconstruct the POc contrast between them.

Nalik (New Ireland)	dua-	REFERENCE	duāk	ADDRESS?	(Chinnery 1929)
Raga (N Vanuatu)	tua-	sG	tuaɣa-	eG	(Taylor 2008)
Ninde (C Vanuatu)	tua-	REFERENCE	tuaʔa	??	(Clark 2009)
Wayan (W Fiji)	tutua	ADDRESS	-tuka	REFERENCE	(Pawley & Sayaba 2022)

POc		*tua-	‘elder s.s. sibling’ (Milke 1958b: *tuqa)
Adm	Mussau	tue-	[{PsG}eC] ‘elder sibling’
NNG	Lukep	tua-	‘elder s.s. sibling’
NNG	Mindiri	tua-n	‘elder s.s. sibling’
NNG	Ulau-Suain	tua-	‘elder s.s. sibling’
NNG	Yabem	tɪwa-	[♂{PsG}eS FBeS] ‘♂elder brother’
PT	Kilivila	tua-	[MZSeC eGE] ‘elder s.s. sibling’ (Malinowski 1929)
PT	Molima	tua-	[{PseG}C] ‘elder s.s. sibling’
MM	Tigak	tua-	‘eldest brother’
MM	Barok	to-	[{PoG}sCsC] ‘s.s. sibling’
MM	Sursurunga	tua-	‘s.s. sibling’
SES	Baegu	(sa)ua-	[{PsG}eC] ‘elder sibling’
SES	Bauro	-uwa	‘♀elder sister’ (ka-uwa eZ, wa-uwa eB)
NCV	Merlav	tuo-	‘elder s.s. sibling’
NCV	Lolovoli	tue-	‘s.s. sibling’
NCV	Raga	tua-	‘s.s. sibling’
NCV	Apma	tua-	‘elder sibling’
NCV	Labo	tua-	[FFF ♂FMBS ♂MFMBS] ‘elder s.s. sibling’
NCV	Uripiv	tua-	[HB] ‘elder s.s. sibling’
PSV		*-tua-	‘elder s.s. sibling’ (Lynch 2001c)
SV	Kwamera	p-rea-	[{PsG}eC] ‘elder s.s. sibling’
SV	Anejom̃	e-twa-	[FBsC] ‘s.s. sibling’
PNCal		*tuka-	‘elder s.s. sibling’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	kʰia-	‘elder s.s. sibling’
NCal	Fwâi	hie-	‘elder s.s. sibling’
NCal	Cèmuhî	cuɔ-	‘elder s.s. sibling’
NCal	Kwênyii	tē-	‘elder s.s. sibling’
Fij	Wayan	tutua	[eG FeBC] ‘elder s.s. sibling’

POc		*tuaka-	‘elder s.s. sibling, elder parallel cousin’ (ACD; Milke 1958b: *tuqaka)
NNG	Tuam	toɣa-	‘elder s.s. sibling’
NNG	Bam	tikua	‘elder s.s. sibling’ (metathesis: POc *tuaka > *takua)
NNG	Manam	toʔa	[PGeC] ‘elder sibling’
NNG	Kis	təkua	‘elder s.s. sibling’ (metathesis: POc *tuaka > *takua)
MM	Bola	tuka-	‘elder s.s. sibling’
MM	Nalik	duāk	‘♀sister’ (Chinnery 1929)
MM	Varisi	toɣa-	[PGseC] ‘elder s.s. sibling’
SES	Bugotu	toɣa-	‘elder s.s. sibling’ (Ivens 1940)
SES	Lengo	toha-	[♂{PoG}eS] ‘♂elder brother’
SES	Kwaio	oʔa-	‘elder brother’
NCV	Raga	tuaɣa-	[MP] ‘elder sibling’
NCV	Araki	roha-	[WZeH HeB] ‘elder brother’
NCV	Vao	toɣa-	[{PsG}eC {HeB}] ‘elder sibling’
NCV	Big Nambas	tuɣa-	[♀FeBD ♀HeBW] ‘♀elder sister’
NCV	Labo	tuaʔa-	‘elder brother’
NCal	Fwâi	hiohã	‘♀elder sister’
NCal	Iaai	tuha-	‘elder s.s. sibling’
NCal	Nengone	tok	‘elder s.s. sibling’
Fij	Wayan	tuka-	[eG FeBC] ‘elder s.s. sibling’
Fij	Bauan	tuaka-	‘elder s.s. sibling’
PPn		*tuaka-, *tuakana	‘elder s.s. sibling’ (Marck 1996)
Pn	Samoan	tuaʔa(a)	‘elder s.s. sibling’ (Pratt 1862)
Pn	Rapanui	tuakana	‘elder brother, elder cousin’ (Churchill 1912)
Pn	Māori	tuaka-na	[PGeC] ‘elder sibling’
Pn	Tahitian	tuaʔana	‘elder sibling’ parent’s elder sibling’s child’
Pn	Hawaiian	(kai)kuaʔana	‘elder sibling’ elder cousin’

With the exception of Bugotu kaka-, a Maringe loan, Oceanic reflexes of *kaka ‘elder s.s. sibling’ are confined to WOc languages, but *kaka can be safely assigned to POc because it is clearly inherited from PMP without change in meaning.

There is also a small cognate set reflecting *kaka but denoting a parent’s o.s. sibling (§2.4.1.2.4). It is not clear how, if at all, this is related to the set below.

PAn		*kaka	‘elder sibling’ (ACD)
PMP		*kaka	‘elder s.s. sibling’ (ACD)
POc		*kaka	‘elder s.s. sibling (ADDR), elder parallel cousin (ADDR)’ (Milke 1958b)
NNG	Gitua	kaka	‘elder s.s. sibling’
NNG	Mindiri	kak	‘elder s.s. sibling’
NNG	Manam	aʔa	‘elder sibling’
NNG	Yabem	kaka	‘elder s.s. sibling’
PT	Sudest	ɣaɣa-	‘s.s. sibling’
PT	Sinaugoro	kaka-	‘elder (s.s.?) sibling’
PT	Western Motu	kaka-	[PGeC] ‘elder s.s. sibling’ (Groves 1958; Lister-Turner & Clark 1954)
PT	Roro	ʔaʔa-	[PeGC] ‘elder sibling’
MM	Babatana	kaka	‘elder sibling’
MM	Kubokota	kaka	‘elder (s.s.?) sibling’
MM	Kokota	kaka	[PP] ‘elder sibling’

cf. also:

SES

Bugotu

kaka

‘elder sibling’ (Maringe loan)(Ivens 1940)

As Marck (1996:223–226) notes and discusses, there are two competing forms for PPn ‘elder s.s. sibling’: *tuaka-, *tuakana (above) and *toqakete (below).The data require the reconstruction of both. PPn *tuaka-, *tuakana must be reconstructed on the basis of external evidence, and the innovatory *toqakete on the basis of the reflexes listed below. Their reflexes give us no clue as to why both terms occurred.

Clark (1980) notes that in non-Eastern Polynesian languages, reflexes of PPn *taqokete have the meaning ‘elder s.s. sibling’, but in EPn languages, this shifts to ‘s.s. sibling-in-law’, replacing PPn *maqā (§2.4.2.2).

PPn		*taqokete	‘elder s.s. sibling’ (Marck 1996)
Proto Tongic		*taqokete	‘elder s.s. sibling’ (Marck 1996)
Pn	Tongan	taʔokete	[{PesG}sC] ‘elder s.s. sibling’ (Völkel 2015)
Pn	Niuean	taokete	‘elder s.s. sibling, elder s.s. cousin’
PNPn		*taqo-kete	‘s.s. sibling’ (Marck 1996)
Pn	Tuvalu	takete	‘s.s. sibling’
Pn	East Futunan	taʔokete	‘s.s. sibling’ (Biggs in POLLEX)
Pn	Rennellese	taʔokete	[{MBeS} FPGSCesC] ‘elder s.s. sibling’
Pn	Pukapukan	taokete	‘s.s. sibling-in-law’
PEPn		*taqokete	‘s.s. sibling-in-law’ (Marck 1996)
Pn	Rapanui	taʔokete	‘sibling-in-law’ (Fuentes 1960)
Pn	Tahitian	tauete	‘s.s. sibling-in-law’
Pn	Marquesan	toete	‘s.s. sibling-in-law, father’s sister’s husband’
Pn	Māori	taokete	‘s.s. sibling-in-law’
Pn	Hawaiian	(kai)koʔeke	‘s.s. sibling-in-law’

4.1.5.3. Opposite-sex sibling

The vast majority of Oceanic languages make a distinction between s.s. siblings and o.s. siblings. A language has either one term for all o.s. siblings, or two terms, one for ‘♀brother’ and one for ‘♂sister’. As the tabulation below shows, the s.s. sibling terms described in the previous two subsections make an implicit statement about the referent’s sex, that it is the same as EGO’s. They may or may not distinguish between younger and older.

	younger/elder?	female/male?
s.s. sibling term	yes or no	same as EGO
Type A o.s. sibling term	mostly no	opposite to EGO
Type B o.s. sibling term	no	opposite to EGO + ALTER’s sex

O.s. sibling terms almost never make a younger/elder distinction. They make an implicit statement about the referent’s sex (Types A and B), and Type B also redundantly specifies ALTER’s sex explicitly. This has the effect that a female EGO uses one term for an o.s. sibling, and a male EGO uses another term.

There is evidence that PMP had Type A o.s. sibling terms, which Blust (1994) reconstructs as *betaw ‘♂sister’, *ñaRa ‘♀brother’. Both terms are reflected in various WMP and CMP languages, but not in EMP (SHWNG and Oceanic). Instead, EMP languages have either Type A terms or they have Type B terms that reflect PMP *maRuqanay ‘male, man, husband’ and *babinahi ‘female, woman, wife’. Their POc reflexes are respectively *mʷaqane and *papine (vol.5:50–55).

This raises two questions. First, how did this come about? And second, how did POc speakers know whether a term referred to someone’s spouse or to someone’s sibling?

Answering the second question first, PEMP and POc distinguished between direct possession, whereby the possessed noun took a suffix indicating the person and number of the possessor, and indirect possession, where the possessor suffix was attached to a separate classifier morpheme (Lichtenberk 1985a). Stress shifted one mora to the right when a single- mora suffix was added.³⁴ The o.s. sibling terms were directly possessed, the spouse terms indirectly. Hence, for example, POc *papiné-gu ‘my sister’ contrasted with *ná-gu papíne ‘my wife’ (where *na- is a classifier).

		male o.s. sibling	female o.s. sibling
wMP:	Sangir	mahuane	bawine
	Tae’	anak-muane	anak-dara
cMP:	Kambera	ana-mini	ana-wini
	Hawu	na mone	na weni
	Leti	nara muani	nara puate
	Paulohi	leu manawa	leu pipina

Non-EMP data discussed by Blust (1994) and tabulated below enable us to answer the first question. These are terms which, Blust argues, did not yet exist in PMP but later replaced reflexes of *betaw and *ñaRa. Except for Sangir, all the tabulated terms have two elements, the first denoting an o.s. sibling, the second the sex of the referent. In Sangir the first element is missing, i.e. it behaves like POc. Of the second elements, all in the ‘male o.s. sibling’ column reflect PMP *maRuqanay, and all except Tae’ and Leti in the ‘female o.s. sibling’ column reflect PMP *babinahi.

Blust provides evidence that these two-element terms were treated as compounds. There was no direct/indirect distinction in wMP and some cMP languages, and possession was direct, i.e. by suffix. In Tae’ (S Sulawesi) there is a contrast between the two constructions below, from van der Veen (1940:17ff.).

• • Tae’ (CMP): (van der Veen 1940: 17)

    ‘my male child = my son’

    anak-ku	muane
    child-my	male

  • Tae’ (CMP): (van der Veen 1940: 17)

    ‘her male o.s. sibling = her brother’

    anakmuane-na
    male.cross.sibling-her

Blust, citing Fischer (1957:5), finds a similar contrast in Kambera between ana-ŋgu mini ‘my son’ and ana.mini-ŋgu ‘my brother’. Because the sibling terms were compounds, the possessor suffix was attached to the second element, which reflected PMP *maRuqanay or *babinahi.

It is easy to see that the POc o.s. sibling terms *mʷaqane- and *papine- are cognate with the second element of the terms tabulated above. The first element dropped out, as it had in Sangir, and the suffix remained. The surviving second element was marked as an o.s. sibling term by its possessor suffix and right-shifted stress.

Although many of the reflexes of POc *papine- and *mʷane- below are glossed ‘o.s. sibling’ one can take this to be an extension of their earlier meanings ‘♂sister’ and ‘♀brother’, that is, ‘female o.s. sibling’ and ‘male o.s. sibling’. If ‘o.s. sibling’ were their prior meaning, one would expect a scattering of ‘♂sister’ and ‘♀brother’ in each set, but this is not what is found.

Only in one subgroup, NNG, does there seem to have been a relatively early shift of a reflex *mʷaqane to the generic sense ‘o.s. sibling’.

PEMP		*babinay	‘♂sister’
CB	Wandamen	vavi[ni]	‘♂sister’
POc		*papine-	‘♂sister’, ♂female parallel cousin’ (Milke 1958b)
MM	Nalik	fafna-	[FBoC ♀FBD WBW] ‘o.s. sibling’ (Chinnery 1929)
MM	Tangga	fefne-	‘o.s. sibling’
MM	Patpatar	hahini-	‘sibling’
MM	Halia	hahina-	[PGsC] ‘o.s. sibling’
MM	Hahon	wevne-	[{PsG}oC] ‘o.s. sibling’
MM	Torau	aine-	‘o.s. sibling’
MM	Varisi	vavani-	[PGoC EsGE] ‘o.s. sibling’
MM	Roviana	vavene-	‘♂sister’ (Capell 1943b)
PSES		*vavine-	‘♂sister’
SES	Bugotu	vavine-	‘o.s. sibling’ (Bogesi 1948)
SES	Lengo	vavine-	[♂{PoG}D] ‘♂sister’
SES	Longgu	vavune-	‘o.s. sibling’ (Hogbin 1938a)
SES	’Are’are	hahone-	‘o.s. sibling’
SES	Arosi	haho-	‘o.s. sibling’
PNCV		*vavine-	‘♂sister’ (Clark 2009)
NCV	Kiai	vavine-	‘♀sister’ (Lynch 2004d)
NCV	Big Nambas	vən-	[♂{PsG}D] ‘♂sister’
NCV	Neve’ei	vivinu-	[♂{PsG}D] ‘♂sister’
NCV	Lewo	vine-	‘♂sister’
PSV		*na-[va]vine-	‘♂sister’ (Lynch 2001c)
SV	Sye	veven, vevne-	‘brother, ♂sister’
SV	Lenakel	no-uinə-	‘♂sister’
Mic	Chuukese	fēfine-	[♂PGD] ‘♂sister’ (Goodenough 1951)
Fij	Rotuman	(sæŋ)vǣvǣne	[♀PGS] ‘♀brother’ (Howard 1970)

cf. also:

SV

Anejom̃

n-ataheñ

‘♂sister’ (< POc *qatapine, vol. 5:56)

PEMP		*muaqanay	‘♀brother’
CB	Wandamen	muani	‘♀brother’
POc		*mʷaqane-	‘♀brother, ♀male parallel cousin’ (?) (Milke 1958b: *［m］ane; Milke 1968)
Adm	Baluan	mʷane-	[♀{PsG}S] ‘♀brother’
NNG	Sio	mane	[{PoG}oC] ‘o.s. sibling’
NNG	Bing	mane	‘o.s. sibling’ (Lincoln 1978)
NNG	Sepa	mone	‘elder sibling’ (Z’graggen 1974a)
NNG	Ali	məne-	‘mother’s brother’
PSES		*mʷai-mʷane	‘o.s. sibling, cross-sex cross-cousin’
SES	To’aba’ita	wai-waena	[PGoC] ‘o.s. sibling’ (fossil -na)
SES	Lau	wae-	‘o.s. sibling’
SES	Kwara’ae	wai	[MZH FZH ♂ZC ♂WBC] ‘mother’s brother’
PNCV		*mʷa(qa)ne	‘♀brother’ (Clark 2009: 152–153) 35
NCV	Akei	mane-	‘♀brother’
NCV	North Ambrym	mʷena-	‘♀brother’ (Guiart 1951)
NCV	Paamese (South)	mone-	‘♀brother’
NCV	Big Nambas	nn̼ana-	‘♀brother’
NCV	Neve’ei	manu-	[♀MZS ♀FBS] ‘♀brother’
NCV	Lewo	mʷene-	‘♀brother’
PSV		*[na]mʷane-	‘♀brother’ (Lynch 2001c)
SV	Sye	man, mano-	‘♀brother’
SV	North Tanna	mʷanə-	‘♀brother’
SV	Lenakel	man-	‘♀brother’

PNCal		*mʷane	‘♀brother’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	mʷãla-	‘elder brother’
NCal	Cèmuhî	mʷane-	‘elder brother’ (archaic; Ozanne-Rivierre 2000)
NCal	Iaai	mañi-	‘o.s. sibling’ (Ray 1917)
PMic		*mʷāne-	‘♀brother’
Mic	Nauruan	mʷæn	‘o.s. sibling’ (Hage and Marck 2002)
Mic	Kiribati	mʷane	[{PsG}oC] ‘o.s. sibling’ (Lambert 1981)
Mic	Marshallese	mmʷahan	[♀HB ♀ZH ♂WB] ‘♀brother’ (Spoehr 1949a)
Mic	Puluwatese	mʷǣne-	‘♀brother’ (Bender et al. 2003a)
Mic	Chuukese	mwǣni-	‘♀brother’ (Bender et al. 2003a)
PCP		*ŋʷaqane-	‘♀brother’
Fij	Wayan	ŋʷane-	[{PsG}oC] ‘o.s. sibling’
Fij	Bauan	ŋane-	‘o.s. sibling’
Fij	Tavuki	ŋane-	‘mother’s brother’

cf. also:

Mic	Satawalese	mʷæne-nnap	‘mother’s brother’
Mic	Carolinian	mʷǣle-llap	‘firstborn son, parent’s eldest brother’
Mic	Woleaian	mʷāle-nnape	‘mother’s brother’(Bender et al. 2003a)
Mic	Ulithian	mal lapa-	‘mother’s brother’

PPn preposed *tua- to its reflexes of the forms above, giving *tua-fafine and *tua-ŋaqane. The *tua- element regularly reflects POc *tuRaŋ ‘friend, companion; relative of ego’s generation’ (pollex; §2.4.1.5.5). Whilst glosses of a few other reflexes of *tuRaŋ mention ‘o.s. sibling’, as many mention ‘s.s. sibling’. It seems that its denotation has narrowed in various ways across languages, including to ‘o.s. sibling’ in PPn.

PPn		*tua-fafine, *tua-fine	‘♂sister’ (Marck 1996)
Pn	Tongan	tuo-fefine	[♂{PG}D] ‘♂sister’ (Völkel 2015)
Pn	Tokelauan	tua-fafine	[♂{PG}D] ‘♂sister’ (Huntsman 1971)
Pn	Rennellese	tua-fafine	[♂{PsG}D ♂F{PsG}CD] ‘♂sister’
Pn	Pileni	tʰuohine	[♂PGD] ‘♂sister’
PEPn		*tua-fine	‘♂sister’ (Marck 1996)
Pn	Māori	tuahine	‘♂sister’
Pn	Tahitian	tuehine	[♀PGD] ‘♂sister’
Pn	Marquesan	tuehine	‘♂sister, father’s sister’
Pn	Hawaiian	kuahine	‘♂sister’

PPn		*tua-ŋaqane	‘♀brother’ (Marck 1996)
Pn	Tongan	tuo-ŋaʔane	[♀{PG}S] ‘♀brother’ (Völkel 2015)
Pn	Samoan	tuaŋane	[♀PGS] ‘♀brother’ (Williamson 1924, vol 2)
Pn	Tokelauan	tuaŋane	[♀MZS] ‘♀brother’ (Macgregor 1937)
Pn	Rennellese	tuŋaʔane	[♀{PsG}S ♀F{PsG}CS] ‘♀brother’
Pn	Pileni	tuoŋane	[♀PGS] ‘♀brother’
Pn	Pukapukan	tuŋāne	‘♀brother’ (from Rarotongan)
Pn	Māori	tuŋane	‘♀brother’
Pn	Tahitian	tuʔane	[♀PGS] ‘♀brother’
Pn	Marquesan	tunane	‘♀brother’ (dialect variants tukane, tuʔane)
Pn	Hawaiian	kunane	‘♀brother’

The Chuukic (Mic) set below appears also to be derived in some way from POc *mʷaqane-. At the end of the *mʷaqane- cognate set above several Chuuukic reflexes are listed under ‘cf. also’. These items give some indication of how PChk *mʷeŋ[ae]ya might have arisen. They are compounds consisting of reflexes of POc *mʷaqane-/PMic *mʷāne- ‘♀brother’ and PMic *lapa ‘big’, literally ‘big brother of a woman’, here meaning ‘mother’s brother’. This opens up the possibility that PChk *mʷeŋ[ae]ya ‘♀brother’ reflects earlier ‘small brother of a woman’ or ‘true brother of woman’, but the etymology of the latter part of the reconstruction, presumably †*ŋ[ae]ya, is unknown.

PChk		*mʷeŋ[ae]ya	‘♀brother’ (Bender et al. 2003a)
Mic	Puluwatese	mʷəŋeya-	[{PsG}oC ♂ZD MZDoC ♀MB] ‘o.s. sibling’
Mic	Chuukese	mʷoŋeya-	‘♀brother’ (Bender et al. 2003a)
Mic	Satawalese	mʷeŋeya-	[PGoC ♂MMZD ♀MMZCS] ‘o.s. sibling’
Mic	Woleaian	mʷaŋeya-	‘o.s. sibling’ (Bender et al. 2003a)
Mic	Sonsorolese	mʷeaŋa-	‘o.s. sibling’ (Capell 1969)
Mic	Ulithian	mʷæŋæ-	‘♀brother’

A distinction was made above between Type B o.s. sibling terms like POc *papine- and *mʷaqane- that distinguish male and female lexically, and Type A o.s. sibling terms that do not make this distinction. There is just one Oceanic Type A term with quite widespread reflexes, and it occurs in two variants, *lopu- and *lipu-. Reflexes are listed below. In some cases assignment to one or the other variant is difficult.

The *lopu- set is small and has an odd distribution: New Guinea Oceanic (NNG + PT) and Kosraean (Mic). Bender et al. (2003a) take Kosraean lɔ, lou- ‘♂sister’ to reflect POc *lopu-. If they are right, then the reconstruction of POc *lopu- is supported. Otherwise it is reconstructable only to PNGOc.

Another small group of possible cognates is given below. Its members are found in Cenderawasih Bay. Unfortunately, knowledge of CB historical phonology is insufficient for us to be certain that these are cognates of POc *lopu-. The initial consonant correspondence regularly reflects PEMP *l-. The loss of *-p- in Ambai, Wooi and Wandamen is expected. Its loss in Irarutu may be irregular. If these prove to be cognates of putative POc *lopu-, then the latter can be reconstructed and so can PEMP *ləpu-.

The *lipu- variant below is reconstructable only to PNGOc. Included here are reflexes of the form lu-, that have lost the vowel of the initial syllable. These may reflect *lopu-. Vitu livuka (under ‘cf. also’) is probably borrowed from a NNG language. It reflects *†lipuq: the regular Vitu reflex would be livu-.

POc		*lopu-	‘o.s. sibling’ (Chowning 1991; Milke 1965)
SJ	Sobei	dafu-	[EoG {PsG}oC EPGoC] ‘o.s. sibling’
NNG	Sera	lo(m)	‘o.s. sibling’
NNG	Adzera	nafu-	[{PsG}oC {PoG}oC] ‘o.s. sibling’
PT	Sudest	louy-	‘o.s. sibling’
PT	Bwaidoga	novu-	[PGoC] ‘o.s. sibling’
PT	Gapapaiwa	novu	‘o.s. sibling’
PT	’Auhelawa	lou	‘o.s. sibling’ (Milke 1965: 345)
PMic		*lowu-	‘♂sister’ (Bender et al. 2003a)
Mic	Kosraean	lɔ, lou-	‘♂sister’

Proto Cenderawasih Bay		*ru(i,o)-	‘o.s. sibling’
CB	Irarutu	rui-	‘o.s. sibling’ (van den Berg and Matsumura 2008)
CB	Ambai	roro-	‘o.s. sibling’
CB	Wooi	ra-ruo-	‘o.s. sibling’ (Sawaki 2016)
CB	Wandamen	ra-ruo	‘o.s. sibling’

PNGOc		*lipu-	‘o.s. sibling’ (Chowning 1991)
NNG	Gitua	livu	‘o.s. sibling’
NNG	Mangap	lu-	‘younger o.s. sibling’
NNG	Tami	lu, liwu-	‘♂sister’
NNG	Maleu	liwa	‘o.s. sibling’ (Hooley 1971)
NNG	Lukep	lui-	‘o.s. sibling’
NNG	Kaulong	e-lu(t)	‘o.s. sibling’
NNG	Lamogai	lu-	‘elder o.s. sibling’ (Ross, fieldnotes)
NNG	Mangseng	lu-	‘♂sister’
NNG	Mengen	liu-	‘o.s. sibling’
NNG	Takia	lu-	‘o.s. sibling, o.s. cross-cousin’
NNG	Wogeo	lu	[FBoC MZoC] ‘o.s. sibling’
NNG	Kairiru	lu-	‘o.s. sibling’
NNG	Yabem	lù	‘♀brother’
NNG	Yabem	lù-ò	‘♂sister, ♂mother’s sister’s daughter’ (-ò FEM)
NNG	Numbami	lu-	‘brother, male cousin’
NNG	Kapin	li	‘o.s. sibling’
NNG	Patep	li	‘sibling, parallel cousin’
PT	Kilivila	lu- -ta	‘o.s. sibling’ (Lawton f.c.)
PT	Muyuw	nu- -t	‘o.s. sibling’
PT	Misima	nu-	‘o.s. sibling’
PT	Gumawana	niu-	‘o.s. sibling, o.s. cross-cousin’

cf. also:

MM

Vitu

livuka

‘o.s. sibling’

Marck (1996) reconstructs PPn *kawe ‘o.s. sibling’, but there is no Tongic or external evidence for it, unless one accepts as external evidence Bauan and Lau Fijian *weka ‘o.s. sibling’ and posits metathesis. Pawley (1981) does not reconstruct this term. Pawley (pers. comm.) points out that its reflexes are restricted to Pn outlier languages, so that even a PNPn reconstruction is insecure.

Pn	Anutan	kave	‘o.s. sibling, cross-sex cousin’
Pn	Tikopia	kave	‘o.s. sibling’
Pn	Ifira-Mele	kave	‘o.s. sibling’
Pn	West Futunan	kave	‘o.s. sibling, father’s sibling’s cross-sex child’(Capell 1958)
Pn	Pileni	kave	‘s.s. cross-cousin’
Pn	Takuu	kave	‘o.s. sibling, o.s. sibling, father’s sibling’s cross-sex child’
Pn	Nukumanu	kave	‘o.s. sibling, cross-sex cousin’
Pn	Luangiua	ʔave	‘o.s. sibling, cross-sex cousin’
Pn	Sikaiana	kave	‘o.s. sibling, cross-sex cousin’

4.1.5.4. Cousins

No dedicated terms for cousins are reconstructable. As a consequence of bifurcate merging (§2.2.1) parallel cousins are treated as s.s. siblings in most Oceanic languages. Some languages have a dedicated term for ‘cross-cousin’, but the few possible reconstructions take account of the data for very few languages (less than five).

4.1.5.5. Companion, friend, close relative

Milke (1968) glossed PMP *tuRaŋ as ‘companion’. His supporting reflexes not only support this gloss, but also denote a mixed bag of kin relationships, which, as the listing below shows, only grows as one casts the net wider. Whereas it is usually possible to attach an unambiguous gloss to the kinship terms discussed in this chapter, in the case of *tuRaŋ this is impossible.

PMP		*tuRaŋ	‘kinsman, relative (undefined)’ (ACD)
POc		*tuRaŋ	‘friend, companion; relative of ego’s generation’ (Milke 1968: ‘companion’; ACD: ‘kinsman, relative (undefined)’)
NNG	Dami	tura	‘o.s. sibling’
NNG	Takia	tura-	‘friend’
PT	Wagawaga	tura	‘friend, spouse’
PT	Motu	tura	‘a friend (of the speaker’s sex)’ (Lister-Turner & Clark 1954)
MM	Bola	tura	‘cross-cousin’ (Goodenough 1997)
MM	Nakanai	tula-	‘co-wife, spouse’s s.s. sibling’
MM	Lavongai	tuŋə	‘older brother’ (archaic); ‘friend’ (Fast and Fast 1989)
MM	Tigak	tiga-	‘s.s. sibling’
MM	Barok	tura-	‘co-husband’

Sense can be made of all this by glossing POc *tuRaŋ ‘friend, companion; relative of ego’s generation’. It is reasonable to assume that friendships between relatives of the same generation were normally between relatives for whom friendship was not tabooed, and that these relations varied somewhat among immediately pre-modern communities.

Table 2.15 Glosses attributed to reflexes of *tuRaŋ

Gloss	Gen.	s/o	Language
friend, companion	?	?	Takia, Wagawaga, Lavongai, Raga, Paamese, Nokuku, Namakir
s.s. friend	?	s	Motu
relative	?	?	Babatana, Roviana, Ughele
G0 relative of speaker’s moiety	?	?	Tolai (Fingleton 1986)
same-generation relative of opposite moiety	0	?	Nehan
sibling	0	?	Merei, Paamese, Namakir
elder sibling	0	?	Lavongai
s.s. sibling	0	s	Tigak, Tolai, Halia, Raga
cross-cousin	0	?	Bola, ’Are’are
o.s. sibling	0	o	Dami, Mota, Pn
parent’s s.s. sibling’s spouse’s child	0	?	Kwaio
spouse	0	o	Wagawaga, Araki
sibling’s spouse	0	?	Lelepa
elder sibling’s spouse	0	?	Neve’ei
spouse’s elder sibling	0	?	Neve’ei
spouse’s s.s. sibling	0	o	Nakanai, Neve’ei
spouse’s o.s. sibling	0	s	Neve’ei
spouse’s s.s. sibling’s spouse	0	s	Nakanai
spouse’s o.s. sibling’s spouse	0	o	To’aba’ita, Kwara’ae
co-wife	0	o	Nakanai
co-husband	0	o	Barok
husbands of two sisters	0	s	Patpatar

4.2. In-laws (affinal kin)⇫

Oceanic languages typically have three in-law categories: spouse, sibling-in-law, and parents- and children-in-law.

4.2.1. Spouses: wife and husband⇫

POc term *qasawa- ‘spouse’ is one of the more widely reflected POc kinship terms, but it is not reflected in SES languages or in Fiji, where its place is taken by PEOc *wati-.

A striking feature of the cognate set supporting *qasawa- is the wide variety of forms to which it has given rise—far more varied than, say, *tama- ‘father’ or *natu- ‘child’.

However, the reasons for this variety are purely phonological. The form *qasawa- had three syllables (four or five when a possessor suffix is added), and has been subject to the Oceanic tendency towards two-syllable roots. The onset consonants of the three syllables are ones that are unstable in Oceanic. Initial *q- is typically either lost, leading sometimes to loss of the first syllable, or is strengthened to k-. Intervocalic *-s- easily becomes -z- or -r-, or -h- and then disappears; in many NNG languages *-s- has become -y-. The syllable *-wa- has often become -wo-, or, as in Proto New Caledonian *-o-.

PAn		*qasawa	‘spouse’ (ACD)
POc		*qasawa-	‘spouse, husband’ (ACD; Milke 1958b: *acawa)
Adm	Aua	harāu	‘wife’
Adm	Baluan	asoa	‘husband’
SJ	Sobei	eso-	‘spouse’
SJ	Bongo	sua	‘husband’
NNG	Tuam	azāwa-, azowa-	‘spouse’
NNG	Mangap	kusi-	‘husband, spouse’
NNG	Sio	kaiwa	‘spouse’
NNG	Maleu	awa	‘husband’
NNG	Mindiri	kiuwa	‘spouse’
NNG	Takia	iwo-	‘spouse’
NNG	Wogeo	yawa-	‘spouse’
NNG	Sissano	awua-	‘spouse’
NNG	Numbami	asowa	‘spouse’
Proto Buang		*rɣa-	‘husband’
NNG	Mapos Buang	rɛɣa-	‘husband’
NNG	Patep	ləya	‘husband’
PT	Sudest	wevo	‘woman, wife’
PT	Kilivila	kʷava	‘wife’ (Lawton f.c.)
PT	Gapapaiwa	kawa	‘spouse’
PT	Sinaugoro	ɣaraɣo-	‘spouse’
PT	Motu	adava	‘spouse, mother’s brother’ (Seligmann 1910:67)
MM	Nakanai	harua	‘husband’
MM	Lavongai	kisŋə	‘spouse’ (Fast and Fast 1989)
MM	Usen Barok	isuo-	‘spouse’
MM	Patpatar	suə-	‘spouse’
PNCV		*asoa-	‘spouse’ (Clark 2009)
NCV	Mota	ra-soa-i	‘spouse’ (ra-: see §2.4.1.2.2)
NCV	Raga	ahoa	‘spouse’
NCV	Araki	rua-	‘spouse’
NCV	Paamese (South)	asō-	‘spouse’
NCV	Port Sandwich	söa-	‘spouse’
NCV	Lewo	o-	‘spouse: husband, wife’
SV	Sye	aso-, ahʷo-	[♀ZH] ‘husband’
PNCal		*qasao-	‘spouse’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	arō-	‘spouse’
NCal	Fwâi	kalō-	‘spouse’
NCal	Xârâcùù	kʷɛ̄tɔ-	‘spouse’
NCal	Iaai	aeã̄-	‘spouse’
PPn		*qahawa-, *qahawana	‘spouse’ (Marck 1996)
PPn		*qahawa(n,ŋ)a	‘marry’
Pn	Tongan	ʔohoana	‘spouse’ (archaic; Churchward 1959)
Pn	Niuean	hoana	‘wife, marry’
Pn	Samoan	āvā	‘wife’ (humble term; Milner 1966)
Pn	Samoan	āvaŋa	‘elope’
Pn	East Futunan	āvaŋa	‘spouse, marriage’
Pn	East Uvean	avaŋa	‘spouse, marriage’
Pn	Tikopia	āvaŋa	‘marry’
Pn	Takuu	āvana	‘marriage, be related as husband and wife’

PEOc		*wati	‘spouse’ (ACD)
PSES		*wati-	‘spouse’
SES	Tolo	ati-	‘spouse’
SES	To’aba’ita	kʷai-	‘spouse’
SES	Kwaio	kʷai-	‘spouse’
SES	Arosi	wai-	‘spouse’
Proto E Fijian		*wati-	‘spouse, cross-sex cross-cousin’
Fij	Tavuki	waci	‘spouse’
Fij	Tokatoka	wati-	‘spouse, cross-sex cross-cousin’
Fij	Moala	wati-	[EoG E{PsG}oC sGEoG sGE{PsG}oC] ‘spouse, s.s. sibling’s spouse, cross-sex cousin’

Did POc have terms for ‘wife’ and ‘husband’? Tentatively, yes, POc *pine meant ‘wife’, but the data do not suggest a corresponding ‘husband’ term.

Oceanic languages tend to use reflexes of POc *papine ‘woman, female’ and *mʷaqane ‘man, male’ for ‘wife’ and ‘husband’. These terms are reconstructed in vol.5(50–55) and discussed in their kinship context in §2.4.1.5.3. Nonetheless, Oceanic languages have sometimes innovated terms that render ‘wife’ and ‘woman’ distinct. At least in one instance, this tendency was already manifest earlier than POc. PMP had a number of forms derived from the root *bahi. It was perhaps originally a stative verb ‘be female’, as many of its derivations include the infix *⟨in⟩, one of several PAn affixes that among other things formed nouns from verbs. In PMP *b⟨in⟩ahi ‘woman, wife’ and *ba-b⟨in⟩ahi ‘female, woman’ are both attested. The latter was ancestral to POc *papine ‘woman, female’. PMP *b⟨in⟩ahi gave rise to PEMP *b⟨in⟩ai and POc *pine ‘woman, wife’, much more sparsely reflected that *papine. The difference in usage seems to have been present at least by PEMP. The cognate set supporting PEMP *b⟨in⟩ai and POc *pine is below.

PAn/PMP		*b⟨in⟩ahi	‘woman, wife’ (ACD)
PEMP		*b⟨in⟩ai	‘woman, wife’
RA	Misool	pin	‘woman, wife’
CB	Ambai	bine	‘wife’
CB	Wandamen	vinie	‘wife’
CB	Warembori	e-vin(do)	‘wife’
POc		*pine	‘woman, wife’ (ACD: ‘female’)
NNG	Barim	vne	‘wife’
NNG	Adzera	fini-	[WZ ♂BW ♂PGSW] ‘wife’
NNG	Mapos Buang	vəne	‘wife’
NNG	Yanta	vni	‘wife’
Proto Solomons Outlier		*fine	‘female relative’
Pn	Ifira-Mele	fine	‘sister-in-law’
Pn	West Futunan	fine	‘wife, woman’
Pn	Pileni	hine	‘daughter’

The cognate set below is a further demonstration of a split that has provided separate ‘wife’ terms. The pre-POc origin of POc *kawe(C) ‘woman’ is unknown. Initial n- occurs in a number of languages where initial *k- is regularly lost and accretion of the POc article *na is attested.³⁶ Final Proto Buang *-h reflects one of POc *-k, *-q and *-R. It is possible that there are chance lookalikes in this set, as the medial consonant is problematic. POc *-w- is reconstructed on the basis of Adm, NCV and NCal languages. The NNG, MM and SES items reflect *-p-.

The NNG terms mean ‘woman’ and contrast with the terms for ‘wife’ above. In Adm languages and Kokota (MM, Santa Isabel) and in SES languages the terms mean ‘wife’ or ‘spouse’.

POc		*kawe(C)	‘woman, wife’ (?) (Lynch 2004d: PSOc *nawe ‘wife’)
PEAd		*n-awe-	‘spouse’
Adm	Lele	n-awe-	‘spouse’ (reciprocal)
Adm	Drehet	n-ewe	‘spouse’
NNG	Yabem	àwɪ	‘woman; wife’
NNG	Wampar	afi	‘woman’
NNG	South Watut	kafe	‘woman’
NNG	Hote	avi	‘woman’
Proto Buang		*avɛh	‘woman’
NNG	Mapos Buang	aveʁ	‘woman’
NNG	Patep	veɣ	‘woman’
MM	Kokota	n-afe-	‘spouse, marriage partner of either sex’
SES	Lau	ʔafe	‘wife, married woman’
SES	Kwara’ae	ʔafe	‘wife’
NCV	Nese	n-au	‘wife’ (Lynch 2004d)
NCV	Big Nambas	n-au-	‘spouse’
NCal	Nemi	nn̥ōe-	‘wife’ (Lynch 2004d)
NCal	Fwâi	nn̥ōe-	‘wife’ (Lynch 2004d)

4.2.2. Siblings-in-law⇫

The English terms ‘sister-in-law’ and ‘brother-in-law’ are ambiguous. A ‘sister-in-law’ is either one’s spouse’s sister or one’s brother’s wife, and a ‘brother-in-law’ is one’s spouse’s brother or one’s sister’s husband. Terms in Oceanic languages sometimes have the same ambiguity, and are accordingly glossed ‘sister-in-law’ or ‘brother-in-law’. In other instances the gloss is more specific, e.g. ‘spouse’s brother’. Often, though, as with siblings, a term indicates the in-law’s sex relative to ego, i.e. ‘s.s. sibling-in-law’ (♀HZ, ♀BW, ♂WB, ♂ZH; more briefly EsG, GsE) or, less frequently, ‘o.s. sibling-in-law’ (♀HB, ♀ZH, ♂WZ, ♂BW; i.e., EoG, GoE).³⁷

The most widely reflected term for a sibling-in-law is POc *ipaR, *ipa-, primarily ‘♀sister- in-law’ and more broadly ‘s.s. sibling-in-law’. The two Kimbe reflexes, Bola and Nakanai, regularly reflect POc *q-, but there is no other evidence for an initial consonant.

PMP		*hipaR	‘sibling-in-law (probably of the same sex only)’ (ACD)
POc		*ipaR, *ipa-	‘s.s. sibling-in-law’ (ACD)
Adm	Mussau	ie-	‘s.s. sibling-in-law’
Adm	Baluan	ipa	‘♀sister-in-law’
SJ	Sobei	ifa-	[PGsCE EPGsC] ‘younger sibling-in-law’
NNG	Tuam	iva-	‘sibling-in-law’
NNG	Mangap	iwa-	‘sibling-in-law’
NNG	Tami	iu, iwa-	‘brother-in-law’
NNG	Takia	iwa-	[{PoG}CsE] ‘spouse’s o.s. sibling’
NNG	Manam	ia	‘sister-in-law’
NNG	Yabem	yàʔ	‘spouse’s brother’
NNG	Numbami	iwa	‘s.s. sibling-in-law’
NNG	Adzera	afaʔ	[♀PGSW HZ] ‘♀sibling’s spouse’
NNG	Hote	ya-	‘spouse’s father, daughter’s husband’
PT	Kilivila	iva- -ta	‘♀sister-in-law’ (Malinowski 1929)
PT	Iduna	ɣiva-	‘♀sister-in-law’
PT	Tawala	iwa-	‘s.s. sibling-in-law’
PT	Sinaugoro	iva-	‘sibling-in-law’
PT	Western Motu	iha-	[PGSW] ‘sibling-in-law’ (Groves 1958)
MM	Bola	ɣiva-	‘s.s. sibling-in-law’
MM	Nakanai	hiva-	‘s.s. sibling-in-law’
MM	East Kara	ifa-	‘father’s sister’s child’
MM	Tangga	ifa-	[♀FZD] ‘♀sister-in-law’
MM	Torau	ia-	‘♀sister-in-law’
MM	Varisi	iva-	‘s.s. sibling-in-law’
MM	Maringe	iva	‘spouse’s sibling’ (Bugotu loan)
PSES		*iva-	‘♀sister-in-law, s.s. sibling-in-law’
SES	Bugotu	iva-	‘sibling-in-law’ (Ivens 1940)
SES	To’aba’ita	θa-ifa-	‘♀husband’s sister’
SES	Kwaio	ifa-	‘sibling-in-law’
SES	Arosi	iha-	‘s.s. sibling-in-law’
SES	Owa	-efa-	‘spouse’s brother’ (wa-efa- EB, ka-efa- EZ)
PNCal		*iva-	‘♀sister-in-law, s.s. sibling-in-law’ (Ozanne-Rivierre 2000)
NCal	Fwâi	ive-	‘♀spouse’s sister’
NCal	Cèmuhî	iɛ-	‘spouse’s sister’
NCal	Dehu	ie	‘spouse’s sibling’
Fij	Nadrogā	iva-	‘♀brother’s wife’
Fij	Wayan	iva-	‘son-in-law’
Fij	Tokatoka	ra-iva-	‘♀sister-in-law’

In PPn the term above was replaced by *maqā.

PPn		*maqā	‘s.s. sibling-in-law’ (Pawley 1981: 284)
Pn	Tongan	maʔa	‘s.s. sibling-in-law’ (Marck 1996)
Pn	Tokelauan	mā	[♂PGDH] ‘sister’s husband’
Pn	Rennellese	maʔā	‘s.s. sibling-in-law’
Pn	Anutan	ma	‘s.s. sibling-in-law’
Pn	Pileni	mā	‘s.s. sibling-in-law’
Pn	Sikaiana	mā	‘spouse’s sibling’

PPn *maqā was in turn displaced in PEPn by the reflex of PPn *taqokete ‘elder s.s. sibling’, which in EPn languages shifted to ‘s.s. sibling-in-law’ (§2.4.1.5.2).

There are seeming gaps in the system of in-law terms. One might expect a specific term for ‘♂brother-in-law’. Oceanic languages often have such a term, but only local reconstructions can be made.

Intriguing is the fact is that there are no NCV reflexes of POc *ipa- and no replacement term is reflected across the archipelago, but there is a possible cognate set for ‘♂ brother-in- law’, stretching across north and central Vanuatu. Clark (2009) reconstructs PNCV *tauwia, but Lynch (2004d) divides the set in two, *taku and *tauwia, to account for the presence or absence of reflexes of *-k-. The division is accepted here, with the rider that, given their similarity of form and identical meaning, the two sets may have a shared history. The form *tau-wia only occurs in Central Vanuatu languages.³⁸ It is clearly a compound. The second element is wia ‘good’ (vol.5:596–597). The first is perhaps reflected in Lewo i-rau and Nguna tā, both glossed ‘friend’ (Clark 2009).

PROc		*taku	‘brother-in-law’ (Lynch 2004d: NCV)
NCV	Akei	tau-	‘♂brother-in-law’
NCV	Araki	rahu-	[HZH] ‘♂brother-in-law’
NCV	Lonwolwol	tovʸa-	‘♂wife’s brother’
NCV	Paamese (South)	tau(letu)	‘♂brother-in-law’
NCV	Vao	tahu-	‘♂brother-in-law’
NCV	Nese	taɣ-	‘brother-in-law’ (Crowley 2006d)
NCV	Tape	e-tɣa-	‘brother-in-law’ (Crowley 2006b)
Fij	Wayan	-daku	‘wife’s sister, husband’s brother’

Proto Central Vanuatu		*tau-wia	‘brother-in-law’ (Clark 2009)
NCV	Atchin	tauwen	[♂DH] ‘husband, sister’s husband’
NCV	Uripiv	tau-	‘♂wife’s brother’
NCV	Port Sandwich	tauia-	‘♂daughter’s husband’
NCV	Nguna	(ana)tawia-	‘♂brother-in-law’
NCV	Lelepa	tawina	‘♂brother-in-law’ (Guiart 1964)

One would also expect a POc term for ‘o.s. sibling-in-law’. A little digging reveals that *taci- ‘(younger) s.s. sibling’ (§2.4.1.5.1) also applied to the spouse of *taci- (sGE). Somewhat counter-intuitively, this means that its secondary sense is ‘o.s. sibling-in-law’, also including ego’s spouse’s siblings of the opposite sex to ego (EoG). There are also a few instances of *tua- ‘elder s.s. sibling’ (§2.4.1.5.2) used in the same way—’few’ because they only occur in languages that make an age distinction among s.s. siblings. In the listings below, glosses that exemplify ‘o.s. sibling-in-law’ are placed before a semicolon, other glosses after it. Mussau tasi- has only the ‘sibling-in-law’ sense, having lost the ‘sibling’ sense.

POc		*taci-	‘o.s. sibling-in-law, younger than ego’; ‘younger s.s. sibling’
Adm	Mussau	tasi-	[sGE] ‘o.s. sibling-in-law’
Adm	Lou	teri-	[esG {PsG}seC] ‘spouse’s s.s. sibling, younger than ego’ (Mead 1934: 342–344)
NNG	Wab	te-u	‘wife’s sister, younger than ego’
NNG	Bing	te-	‘wife’s sister, younger than ego’
NNG	Takia	tei-	[EoG sGE] ‘o.s. sibling-in-law’
NNG	Yabem	lasi-ò	[♂yZ {PoG}yD WFGyD EMByD ♀WMZyD] ‘wife’s sister, younger than ego’
NNG	Mapos Buang	ari-maluʁ	[(♀?)PGCH] ‘♀brother-in-law’ (maluʁ ‘male’)
NNG	Mapos Buang	ari-aveʁ	[(♂?)PGCW] ‘♂sister-in-law’ (aveh ‘female’)
PT	Dobu	tasi-	[{PsG}sC GsE {PsG}sCE {PoG}sCE] ‘s.s. sibling’s spouse’
MM	Bali-Vitu	tazi	[{PoG}sC] ‘wife’s sister; elder s.s. sibling’
NCV	Araki	ve-rasi-	[yZ WByW] ‘brother’s wife, younger than ego’
NCV	Vao	tehi-	[yG {PsG}yC ♂SSS] ‘♂younger brother’s wife, ♀husband’s brother, younger than ego’
Pn	Tokelauan	taina	[sG {PsG}sC {PoG}sC EoGE] ‘s.s. sibling’s spouse’ (Huntsman 1971)
Pn	Tuvalu	taina	[EoG EoGE] ‘spouse’s s.s. sibling’
Pn	Anutan	taina	[sG PGsC] ‘s.s. sibling’s spouse’
Pn	Tikopia	taina	[sG] ‘s.s. sibling’s spouse’
Pn	Nukumanu	taina	‘spouse’s s.s. sibling’
Pn	Luangiua	kaiŋa	[sG PGsC EoGE] ‘o.s. sibling-in-law’

Whilst Lou tio- below pairs with teri- above, Kilivila tua- below pairs with nothing above, because Kilivila does not reflect *taci- but has replaced it with bʷada- ‘younger s.s. sibling, younger o.s. sibling’s spouse, ♂wife’s younger sister’.

POc		*tua-	‘o.s. sibling-in-law, older than ego’; ‘elder s.s. sibling’ (Milke 1958b: *tuqa)
Adm	Lou	tio-	[{PsG}seC] ‘spouse’s s.s. sibling, older than ego; elder s.s. sibling’ (Mead 1934: 342–344)
PT	Kilivila	tua-	[MZSeC eGE] ‘elder o.s. sibling’s spouse’ (Lounsbury 1965; Malinowski 1929)

4.2.3. Parents- and children-in-law⇫

Many Oceanic languages class parents-in-law and children-in-law together, and terms are used reciprocally. Perhaps because this does not distinguish between generations, languages quite often extend a consanguineal term to include an in-law category. Thus the term for a blood-relative is extended to include the equivalent relative of one’s spouse, so that reflexes of POc *tama- ‘father’ and *tina- ‘mother’ are also used for one’s father-in-law and mother-in-law respectively in Yapese, Manam, Chuukic languages and Rennellese (as are the corresponding terms in many English-speaking communities). Similarly, reflexes of POc *natu- ‘child’ are sometimes used for a child’s spouse, for example in Raga and Chuukese. Reflexes of POc *ipa- ‘s.s. sibling-in-law’ are extended to include other in-law generations in a number of languages along the north coast of New Guinea (perhaps a result of contact). Reflexes of PPn *matuqa ‘parent’ are extended to parent-in-law in Anuta, Tikopia and Tongareva.

Sometimes a language has a single term for ‘mother’s brother’ and ‘♂father-in-law’, due originally to asymmetric cross-cousin marriage (§2.4.1.2.5). A more complex example diachronically is found in southern New Caledonia. In Drubea and Kwênyii tũ̃-, reflecting POc *tubu- ‘grandparent’ denotes ‘mother’s brother’ and ‘♂father-in-law’. This reflects two extensions. First, *tubu- reflexes were extended in meaning to include MB (§2.4.1.2.5). Then the term for MB was extended to EF.

These extensions are found in the midst of reflexes of two terms, POc *rawa- and PEOc *puŋao- ‘parent- or child-in-law’. Reflexes of POc *rawa- are found throughout most of WOc, with just a few popping up in NCV and NCal. Reflexes of *puŋao- are found throughout EOc. Thus reflexes of *rawa- and *puŋao- are almost in complementary distribution.³⁹

There is little doubt that *rawa- was the POc term for ‘parent- or child-in-law’,⁴⁰ both because it has a sprinkling of reflexes in SOc and because there is a CB cognate.

PEMP		*r(a,e)wa	‘parent- or child-in-law’
CB	Wandamen	rewa	‘son’s wife’
POc		*rawa	‘parent- or child-in-law’ (Milke 1965: PNGOc)
SJ	Sobei	dawo-	[EPG EPGE CE] ‘parent-in-law’
NNG	Tuam	rawa-	‘parent-in-law’
NNG	Mangap	rwo-	‘parent- or child-in-law’
NNG	Sio	lawa	‘child-in-law’
NNG	Tami	lau, laua-	‘parent- or child-in-law’
NNG	Lukep	rō-	‘parent-in-law’
NNG	Mengen	lao-	‘parent- or child-in-law’
NNG	Takia	rao-	‘parent- or child-in-law’
NNG	Manam	rawa	[E{PsG}] ‘parent-in-law’
NNG	Yabem	lawa-	‘father- or son-in-law, male affines of father’s and great-grandfather’s generations’
NNG	Yabem	lawa-ò	‘mother- or daughter-in-law, female affines of mother’s and great-grandmother’s generations’
NNG	Numbami	lawa-	‘father-in-law, daughter’s husband’
PT	Misima	yawa-	‘father-in-law, daughter’s husband’
PT	Kilivila	yawa-	[sGCE] ‘parent- or child-in-law’ (Lawton f.c.)
PT	Dobu	lawa-	‘mother-in-law, son’s wife’
PT	Iduna	lawa-	‘wife’s parent, child-in-law’
PT	Gapapaiwa	rawa	‘parent- or child-in-law’
PT	Tawala	lago-	‘cross-sex parent-in-law’
PT	Tubetube	rawa	‘spouse’s sibling’
PT	Western Motu	rava-	[E{PsG} GEC EPP CCE] ‘parent- or child-in-law’ (Groves 1958)
MM	Vitu	rava-	‘parent- or child-in-law’
MM	Bulu	lava	‘parent- or child-in-law’
MM	Bola	lavo	‘parent-in-law’
MM	Nakanai	loa	‘wife’s parent, son-in-law’
MM	Tinputz	noa-	‘mother-in-law’
MM	Torau	roa-	[ZH] ‘wife’s parent, daughter’s husband’
MM	Varisi	rava-	[GCE] ‘parent- or child-in-law’
MM	Roviana	roa-	‘parent-in-law’ (Capell 1943b)
NCV	Big Nambas	rawa-	‘parent-in-law’
NCV	Mota	rowoa-	[♀ZHS ♀HGS] ‘♀sister’s husband’ (Vienne 1984)
NCV	Vao	rava-	‘♀sister-in-law’
NCal	Paicî	tʌ̄̃	‘parent- or child-in-law’

The term *puŋao- almost replaced *rawa- in EOc languages, although it is rare in NCV languages. In PPn it split into the expected reflex *fuŋao-na, but with its meaning restricted to ‘child-in-law’, and PPn *fuŋao-ai ‘parent-in-law’.

PEOc		*puŋao-	‘parent- or child-in-law’ (Chowning 1991; Milke 1958b: POc *puŋo)
PSES		*vuŋao-	‘parent- or child-in-law’
SES	Bugotu	vuŋao	‘parent-in-law’ (Bogesi 1948)
SES	Lengo	vuŋau-	[♂GCE ♂FZCCE] ‘parent-in-law, ♂child-in-law’
SES	To’aba’ita	fuŋao	[GDH ZH PZH] ‘daughter’s husband’ (appears to mean ‘male in-law of any generation’)
SES	Kwara’ae	fuŋa	‘parent- or child-in-law’
SES	Arosi	huŋo-	‘parent- or child-in-law’
NCV	Araki	vuŋŋo-	[MBW] ‘mother-in-law’
Proto North New Caledonia		*pmʷao-	‘parent- or child-in-law’ (Ozanne-Rivierre 2000)
NCal	Nêlêmwa	mō-	‘parent- or child-in-law’
NCal	Fwâi	fʷã̄-	‘parent- or child-in-law’
NCal	Cèmuhî	mʷā-	‘parent- or child-in-law’
NCal	Iaai	ũŋo-	‘parent- or child-in-law’
Fij	Wayan	vuŋa-	‘mother- or daughter-in-law’
Fij	Tokatoka	vuŋo-	‘parent- or child-in-law’
PPn		*fuŋao, *fuŋaona	‘child-in-law’ (Marck 1996)
Pn	Niuean	fiŋona	‘child-in-law’
Pn	Tuvalu	fuŋaona	‘child-in-law’ (archaic)
Pn	Rennellese	huŋa	[BDH BSW ZSW] ‘parent- or child-in-law’
Pn	Anutan	puŋona	‘parent- or child-in-law’
Pn	Tikopia	foŋona	‘child-in-law’
Pn	Pileni	huŋo-	‘child-in-law’
Pn	Takuu	hinaona	[sGsCE {GoC}E] ‘parent- or child-in-law’
Pn	Pukapukan	unaoŋa	‘child-in-law’
Pn	Rapanui	hunoŋa	‘child-in-law’ (POLLEX)
Pn	Māori	hunaoŋa	‘child-in-law’ (metathesis)
Pn	Tahitian	hunoʔa	‘child-in-law’ (metathesis)
Pn	Hawaiian	hūnōna	[EGC] ‘child-in-law’

PPn		*fuŋao-ai	‘parent-in-law’ (Marck 1996; Pawley 1981: 284)
Pn	Niuean	(matua) fuŋavai	‘parent-in-law’ (matua ‘parent’)
Pn	Rennellese	huŋabai	‘parent-in-law’
Pn	Pileni	ŋʰovae	‘parent-in-law’
Pn	Māori	huŋarei	‘parent-in-law’ (-r- for †-v-)

Neither *rawa- nor *puŋao- is reflected in Adm languages. Instead, the term for ‘parent- or child-in-law’ is *(ñ,n)ana, identical in form with the POc address term for ‘mother’ (§2.4.1.2.2). Whether there is a historical connection between the two terms is not clear: this may be a chance resemblance. The gloss ‘parent- or child-in-law’ is shown with a question-mark because none of the reflexes has this range of meaning.

PAdm		*(ñ,n)ana	‘parent- or child-in-law’ (?)
Adm	Baluan	nana-	‘parent-in-law’
Adm	Lou	nana-	‘father- or son-in-law’
Adm	Lele	nono	[MMBDH EFZH] ‘father- or son-in-law’ (Mead 1934: 345–347)
Adm	Loniu	ñana	‘father- or son-in-law’
Adm	Nyindrou	ñana-	‘father- or child-in-law’

The fact that neither *rawa- nor *puŋao- has many reflexes in NCV languages was noted above. Instead, there are reflexes of PNCV *bʷalika ‘parent- or child-in-law’. One apparent external cognate is found, To’aba’ita kʷaliʔa-, listed under ‘cf.also’ because its gloss, although a kinship term, is quite different from those of its NCV cognates.

PNCV		*bʷalika	‘parent- or child-in-law’ (Clark 2009: ‘affine’)
NCV	Loh	kʷiliga	[♀ZDH ♀HFB] ‘♀father- or son-in-law’
NCV	Mota	kʷaliɣa	[WPG ZCE] ‘♂parent- or child-in-law’ (Vienne 1984)
NCV	Raga	bʷaliɣa	‘♂father- or son-in-law’
NCV	Kiai	palia-	‘father-in-law, son-in-law’
NCV	Kiai	ve-palia-	‘mother-in-law’ (ve- FEM)
NCV	Araki	paliha-	‘♂father- or son-in-law’
NCV	Araki	ve-paliha-	‘♂mother-in-law’ (ve- FEM)
NCV	Big Nambas	ð̼lia-	‘mother-in-law’
NCV	Uripiv	pʷile-	‘father or mother-in-law’
NCV	Naman	balɣə-	‘wife’s father’
NCV	Port Sandwich	vilax	‘daughter-in-law’

cf. also:

SES

To’aba’ita

kʷaliʔa-

‘member of grandparent or grandchild generation and beyond’

5. Marriage⇫

There are a number of generalisations about Oceanic marriage that emerge from dictionaries and ethnographies. Early Oceanic societies probably had moieties, and a person was constrained to marry someone from the opposite moiety. Often there was a preference for that person to be a cross-cousin. Sister exchange was also evidently practised. A number of languages have a verb meaning ‘marry out of order’, i.e. marry someone of one’s own moiety or division. If a marriage was deemed proper, then the man’s family would pay a negotiated brideprice for his bride. Many languages have a term for ‘co-wife’, pointing to polygyny.

Reconstructable terms are disappointingly few, perhaps because these matters were referred to by idioms. ‘Shake hands’ and ‘take a woman’ are quite widespread Oceanic idioms for ‘marry’. There are, however, two reconstructable verbs for ‘marry’.

The first of these has POc *qasawa ‘spouse’ as its root (cf §2.4.2.1). It is evident from ACD data that cognate verbs also occur in a number of non-Oceanic languages. However, the reconstruction is questionable, as it is plausible that such a term was invented on various occasions during the spread of Oceanic languages into the Pacific. Indeed, the Pn terms below retain the PPn suffix *-na which occurred on inalienably possessed kinship nouns.

POc		*qasawa	‘marry’ (?)
Adm	Lou	asɔ	‘marry’
Adm	Nyindrou	esou	‘marry’
PT	Ubir	yawa-n	‘marry’
NCal	Fwâi	halō-n	‘marry’
NCal	Xârâcùù	xɔyɔ	‘get married’
NCal	Iaai	hɔiɔ	‘get married’
Pn	Niuean	hoa-na	‘wife; marry’
Pn	Samoan	ava-ŋa	‘(woman) marry’
Pn	East Uvean	ʔohoa-na	‘wife; marry’

A more probable reconstruction is POc *laki or *tau-laki ‘marry’. Its origin was perhaps PMP laki ‘male, masculine, man’ (ACD), and occasional reflexes (e.g. Tiang (MM) lek ‘husband’) point to POc *laki ‘married man’. However, compound forms like Tolai (MM) tau-lai ‘married person’ (cf *tau, vol.5, §2.2.1.1) suggest that POc *laki was a stative verb. Either way, reflexes of both *laki and *tau-laki are widely enough reflected as verbs to allow their POc reconstruction as ‘marry, get married’.

POc		*[tau]laki	‘marry; married person’41
Adm	Mussau	laai	‘marry’
PT	Gumawana	nai	‘marry (s.o.)’
PT	Dawawa	naɣi	‘marriage’
MM	Vitu	laki-a	‘married person’ (-a NOM)
MM	Nakanai	tau-lai-la	‘marriage ceremony’ (-la NOM)
MM	Tiang	lek	‘husband’
MM	Madak	(at)lok	‘husband’
MM	Patpatar	tole	‘marry’
MM	Patpatar	t⟨in⟩ole-n	‘wedding; marriage’ (⟨in⟩ NOM)
MM	Tolai	taulai	‘married person’
MM	Ramoaaina	taula	‘marry’
MM	Ramoaaina	t⟨in⟩aula	‘marriage’ (⟨in⟩ NOM)
MM	Siar	taulai	‘marry’
MM	Nehan	le	‘marry, married person’
MM	Banoni	nai	‘marry’
MM	Maringe	tolaɣi	‘marry’
SES	Bugotu	taulaɣi	‘marry’
SES	Lengo	taulaɣi	‘marry’
SES	Longgu	taulai	‘marry’
SES	Bauro	auragi	‘marry’
SES	Arosi	ragi	‘marry’
NCV	Mota	laγ	‘marry, be married’
NCV	Ambae	laki	‘marry, be married’
NCV	Raga	laγi	‘marry, be married’
NCV	Apma	laγ	‘marry, be married’
NCV	Araki	laɣi	‘marry, be married’
NCV	Nokuku	(te)leki-a	‘marry, be married’
NCV	Lonwolwol	le	‘marry, be married’
NCV	Larëvat	ne-laγ	‘husband’
NCV	Lewo	la	‘marry, be married (of woman)’
NCV	Tape	laγ	‘marry’
NCV	Tape	e-laγ	‘husband’
NCV	Port Sandwich	laγ	‘marry, be married’
NCV	Nakanamanga	laki	‘marry, be married (of woman)’
NCV	South Efate	lak	‘be married’
SV	Sye	(empγu)laγ	‘marry, be married’ (empγu ‘dance’)
SV	Anejom̃	(asan)lai	‘marry, be married’

6. Conclusion(s)⇫

The kinship terms we reconstruct are summarised in Table 2.11 and plotted as a tree diagram in Figure 2 in §2.3.5. The research reported here is founded in large degree on the work of others (§2.1), but it is innovative in three respects. First, it is based on a much larger data corpus than previous reconstructions. Second, it modifies the inventory of reconstructed POc kinship terms, adding to and subtracting from the inventory and suggesting resolutions to some disagreements. Third, it examines the structure of the POc kinship terminology in some detail and makes some novel findings.

The inventory of POc terms includes the addition or confirmation of alternative or address terms that Blust (1979) reconstructs for PMP or Chowning (1991) for POc. A summary of these forms is given in §2.4.1.1. They include POc *ama ‘father, father’s brother’, *mama ‘father, father’s brother (ADDR)’, *tata ‘father, father’s brother, other senior males (ADDR)’, *ina ‘mother, mother’s sister’, *ñaña ‘mother, mother’s sister’, *nai ‘mother (ADDR)’, *[bu]bui ‘grandparent, grandchild’. The multiplication of terms for ‘father’ and ‘mother’ is unsurprising when one considers the number of English expressions for them. Also confirmed are Milke’s (1965) reconstructions of *wawa ‘mother’s brother’ and *rawa- ‘parent- or child-in-law’, which are promoted from PNGOc to POc.

A formal novelty is the replacement of Milke’s (1958b) *tuqaka ‘s.s. older sibling’ by two forms, *tua- and *tuaka- (§2.4.1.5.3). An addition is POc *bawa[-] ‘great-great-grandparent, great-great-grandchild’ (§2.4.1.4.3)

It was originally hoped to provide a history of developments in the structure of Oceanic kinship terminologies from POc through to PPn, but this was abandoned for reasons of space. Instead, §2.3 and its subsections provide a typology of Oceanic terminologies, from which something of their history can be gleaned. The typology situates the POc terminology and its descendants in relation to some of the concerns of kinship theorists. It shows that POc had a bifurcate merging terminology, not surprisingly, given that this structure has been attributed to many of its descendants (§2.3.1). Unexpectedly, however, the typology led us to the conclusion that POc was generationally skewed (a ‘Crow’ terminology), or, more probably, that this was a two-state terminology like Dobu or like Fanti of Ghana, with different sets of terms being used for ‘cross’ relations according to circumstances. A fair quantity of data supports the predictions that generational skewing entails (§2.3.6).

A major reason for suspecting that POc had a generationally skewed terminology is the difficulty that others before us have had in reconstructing a term for father’s sister, a relationship that is a ‘converse’ of mother’s brother, for which dedicated terms are ubiquitous in Oceanic languages. The improbability that POc *aya meant ‘father’s sister’ is discussed in §2.4.1.2.6.

Another term that must also be removed from the POc terminology is *puŋao- ‘parent- or child-in-law’. It has no Admiralties or WOc reflexes and is evidently a PEOc innovation, replacing POc *rawa- (§2.4.2.3).

7. Appendix to chapter 2: Ethnographic sources⇫

• Adzera: H. Holzknecht 1976; S. Holzknecht 1986
• Ajië: Leenhardt 1930, 1935; de la Fontinelle 1971
• Akei: Guiart 1958
• Ambae, NE (Lombaha): Allen 1964a
• Ambae, NE (Longana): Lovell 1980
• Ambel: Arnold 2018
• Ambrym, N: Deacon 1927; Franjieh 2012; Guiart 1951; Lane & Lane 1956; Löffler 1960; Scheffler 1970
• Anejom: Lynch 2005
• Aniwa: Capell 1958
• Anuta: Feinberg 1979
• Araki: François 2002
• Arosi, E: Fox 1924
• Avau: Hiroko Sato, fieldnotes
• Avava (Niviar ): Deacon 1934
• Babatana: Money n.d.
• Baegu: H. Ross 1973
• Baluan: Schokkin & Otto 2017
• Bariai: Gallagher 2008
• Barok, Karu: Chinnery 1929
• Barok, Usen: Wagner 1986
• Bauro: Fox 1924
• Belep: McCracken 2012
• Big Nambas: Guiart 1952; Dodd 2014
• Bipi: Mead 1934
• Birao: Hogbin 1937
• Bugotu: Bogesi 1948
• Bukawa: Eckermann 2007
• Butmas: Guiart 1958
• Bwaidoga: Jenness & Ballantyne 1920
• Cèmuhî: Bensa & Rivierre 1982
• Chuukese: Goodenough 1951
• Daakaka: von Prince, 2012, 2017
• Dobu: Fortune 1932
• Drehu: Ozanne-Rivierre 2000
• Drubea: Ozanne-Rivierre 2000
• Fijian, E (Bauan): Capell & Lester 1945-46
• Fijian, W (Deuba): Geddes 1945
• Fijian, E (Lakeba I.): Pauwels 2015
• Fijian, E (Matailobau) Turner 1991
• Fijian, E (Moala): Sahlins 1962
• Fijian, E (Nasau): Cayrol 2015
• Fijian, E (Tokatoka): Nayacakalou 1955
• Fijian, W (Nadi): Capell & Lester 1945-46
• Fijian, W (Nadrogā): Capell & Lester 1945-46
• Fijian, W (Tavuki): Capell & Lester 1945-46
• Fijian, W (Vuda): Capell & Lester 1945-46
• Fijian, W (Wayan): Pawley & Sayaba 2022
• Futuna, E: Burrows 1936
• Futuna, W: Capell 1958; Dougherty 1983
• Fwâi: Ozanne-Rivierre 2000
• Galea: Digim’Rina 1995
• Gedaged: Hannemann 1942
• Gela: Codrington 1891; Hogbin 1938a; Fox 1955
• Guadalcanal, W: Hogbin 1964a
• Hahon: Blackwood 1935
• Hawaiian: Handy & Pukui 1951
• Iaai: Ray 1917; Ozanne-Rivierre 2000
• Ifira-Mele: Clark 1998
• Kapingamarangi: Lieber & Dikepa 1974
• Kara, W: Küchler 2002
• Karavar: Errington 1974
• Kaulong: Goodale 1979
• Kilivila: Malinowski 1929; Lounsbury 1965; Lawton f.c.
• Kiribati: Spoehr 1949b
• Kiribati (Butaritari-Makin): Lambert 1981
• Kove: Chowning 2009
• Kubokota: McDougall 2000
• Kwaio: Keesing 1968
• Kwamera: Lindstrom 1981, 1986, pers. comm.
• Kwara’ae: Deck 1934
• Kwênyii: Ozanne-Rivierre 2000
• Lamusong: Clay 1977
• Larëvat: Deacon 1934
• Lele: Boettger 2015
• Lelepa: Guiart 1964
• Lengo: Hogbin 1964a
• Loh-Toga: Durrad 1940; Mondragón 2014
• Longgu: Hogbin 1964a
• Luangiua (= Ontong Java): Hogbin 1931
• Malua Bay: Deacon 1934
• Manam: Wedgwood 1934
• Mangaia: Hīroa 1934
• Manihiki-Rakahanga: Hīroa 1932b
• Māori: Hīroa 1949
• Mapos Buang: Rambok & Hooley 2010
• Maringe: Boswell 2018
• Marovo: Capell 1943b
• Marquesan: Handy 1923
• Marshallese: Spoehr 1949a, 1949b
• Marshallese (Namu): Pollock 1970
• Maw: Guiart 1964
• Mekeo, N: Mosko 1985
• Molima: Chowning n.d.
• Mota: Codrington 1891; Codrington & Palmer 1896; Vienne 1984
• Motu, W: Groves 1958, 1963
• Mussau: Brownie & Brownie 2007; John Brownie, pers. comm. 2020
• Mwerlap: Durand 2013
• Naha’ai : Wilemp in Deacon 1934
• Nakanai: Chowning & Goodenough 2016
• Nalik: Chinnery 1929, Volker 2020
• Namakir: Sperlich 2019
• Naman: Lagalag in Deacon 1934
• Nauruan: Wedgwood 1936
• Nduindui: Allen 1964b
• Nengone: Ray 1917
• Nehan: Nachman 1978; Glennon & Glennon 2006
• Nêlêmwa: Ozanne-Rivierre 2000
• Nemi: Ozanne-Rivierre 2000
• Neve’ei: Lambumbu in Deacon 1934
• Nguna: Guiart 1964; Facey 1989
• Ninde: Mewun in Deacon 1934
• Notsi: Powdermaker 1933
• Nukumanu: Feinberg 2009
• Nukuoro: Carroll & Soulik 1973
• Owa: Mellow 2014
• Paicî: Leblic 2005
• Pak: Mead 1934:347–349
• Patpatar: Peekel 1908
• Petats: Blackwood 1935
• Pileni (= Vaeakau-Taumako) Feinberg 2013, Næss & Hovdhaugen 2011
• Pingelapese: Schneider 1980
• Pukapukan: Beaglehole 1991; Salisbury & Salisbury 2018
• Puluwat: Elbert 1972, 1974
• Raga: Taylor 2008
• Rapa: Hanson 1970
• Rarotongan: Buse & Taringa 1996
• Rennellese: Elbert & Monberg 1955
• Roro: Monsell-Davis 1981
• Roviana: Capell 1943b
• Sa’a: Ivens 1927
• Samoan: Williamson 1924; Holmes 1957
• Satawalese: Sudo 1985
• Sikaiana: Donner 2020
• Sinesip: Seniang in Deacon 1934
• Sio: Groves 1934
• Sobei: Sterner 1992
• Sori: Romanucci-Ross 1985
• Sudest: Lepowsky 1991
• Sursurunga: Samson et al. 2020
• Sye: Humphreys 1926; Crowley 2000
• Tahitian: Handy 1930
• Takia: Malcolm Ross’ fieldnotes
• Takuu: Moyle 2011
• Tamabo: Rubinstein 1979
• Tangga: Bell 1977
• Teop: Schwartz et al. 2007
• Tikopia: Firth 1922
• Titan: Mead 1934
• To’aba’ita: Lichtenberk 2008
• Tokelau: Macgregor 1937; Huntsman 1971
• Tolai, dialect unknown; not Matupit: Trevitt 1940
• Tolai (Matupit): Fingleton 1986; Simet 1991
• Tongan: Völkel 2015; Douaire-Marsaudon 2015
• Tongarevan (= Penrhyn): Hīroa 1932a
• Torau: Bethwyn Evans’ fieldnotes
• Tuam: Bugenhagen 2011
• Tubetube: Seligman 1910
• Tuvalu: Besnier 2000:626–631
• Ulawa: Ivens 1927
• Ulithian: Lessa 1966; Hage & Marck 2002
• Ura: Lynch 1983c
• Uripiv: Deacon 1934
• Vao: Layard 1942
• Varisi: Scheffler 1965
• Vitu: Blythe 1978
• Vurës: Malau 2016
• Wagawaga: Seligmann 1910
• Wala: Hogbin 1939
• Wandamen: Flaming 1983
• Warembori: Donohue 1999
• Watut, Middle: Fischer 1963
• Whitesands: Humphreys 1926; Casson & Gregory 1976
• Wogeo: Hogbin 1964b, 1970
• Woleaian: Burrows & Spiro 1957
• Wuvulu: Hafford 2014
• Xârâcùù: Ozanne-Rivierre 2000
• Yabem: Streicher 1982
• Yapese: Schneider 1953

Notes⇫