Much of this chapter is reproduced from Lichtenberk (1994), reordered and slightly
revised in the light of more stringent subgrouping assumptions.1 To that work have been added a
number of terms related to fire and its accompaniments. Organization of
the chapter is (1) fire and the fireplace; (2) cooking methods; (3)
preservation and (4) food processing.
Fire was produced throughout the Oceanic world by friction. Often
this was achieved by rubbing a piece of wood rapidly to and fro along a
long groove or furrow cut into a second piece of wood. We have a term
specifically for the action of lighting a fire with fire-plough. Note
that the Central Pacific terms have unexplained vowel fronting.
Figure 19: POc
*suka, *suka-i ‘make fire with
fire-plough’
A second term is POc *usu(q,p), from PMP *usuq
whose primary meaning is probably ‘rub abrasively’, but which in a
number of Oceanic languages is recorded as meaning ‘rub to make
fire’.
Biggs (1993) has reconstructed PPn *kau-natu, for the
implement used. Although this compound is derived from POc elements (POc
*kaiu ‘tree, wood, stick’ and POc *natu ‘child’), we
have no proof that the compound itself is POc. We can be confident,
however, that the term is older than PPn, given that POc *kaiu
was replaced by PPn *raqakau ‘wood, tree’, and POc
*natu by PPn *tama ‘child’.
A separate term for the upper frictioning stick can be reconstructed,
but only for the Central Eastern subgroup of Polynesia:
PCEPn
*kau-lima
‘upper frictioning stick of fire-plough’ (‘stick’ +
‘hand’)
A Papuan Tip language, Minaveha, likewise uses a compound term for
the upper frictioning stick which includes the term for ‘hand’,
nima-kabi (nima ‘hand’ + kabi ‘?’). In at
least two places, Sissano (NNG) and Nakanai (MM), the upper fire stick
is referred to by the term for penis (etin in Sissano,
huti in Nakanai), both reflexes of POc *quti(n).
Across the Oceanic region it is possible to find some variation in
fire-making methods, although we have no reconstructions for methods
other than by fire-plough. Bing on the PNG north coast and Mapos Buang
in the Huon Gulf area (both NNG) offer evidence for a method in which
rattan or vine is sawn rapidly around a piece of wood:
‘make a fire by the traditional method using a piece of rattan,
sawing it back and forth around a piece of dry wood with a split in it,
lying on a little pile of tinder’
Another north coast language, Gedaged, has a term, wol,
referring to a bow string. Fire is produced by pulling a bow string
rapidly back and forth over a piece of wood. The action implied may be
the same as in Carolinian bʷuru-bʷur ‘make a fire by generating
friction with a traditional drill’ (from POc *buru ‘bore a
hole, drill’ ; see Ch. 9,
§4.2).
In addition to the general term POc *kaiu ‘wood, stick’, we
can reconstruct a PEOc term *papia, specifically for firewood.
Note that an identical term is reconstructable for PEOc (cognates in SES
and Fiji) with the meaning ‘cook in native oven’ (§3.1).
Green and Pawley (Ch. 3,
§3.6) note a general term for fireplace or hearth, PMP
*dapuR, POc *rapu(R), with a frequent secondary
meaning, particularly in Oceania, of ‘ashes’. Compounding may be used to
distinguish the fireplace as ‘place of ashes’.
There is a second term, PAn *qabu, POc *qapu, with
similar reference. Blust (ACD) comments: “*qabu is among the
most widespread and stable morphemes in the Austronesian family. It
referred to ashes, and prototypically to the ashes in the fireplace
(hence its recurrent replacement of PMP *dapuR in the meaning
‘hearth’).”
We would expect to be able to reconstruct a term for a trivet, an
arrangement of three stones on which the cooking pot is placed, but no
clear reflex of PMP *dalikan, ‘trivet’ (Blust 1986) has been found, in spite of
the presence of terms for trivet in north-west Melanesia and Fiji. For
example:
A far more soundly based reconstruction is POc *qumun ‘oven
made with hot stones, cook in a stone or earth oven’, which has reflexes
in every major Oceanic subgroup, together with some external cognates.2 However, archaeological evidence
points to such ovens having much greater antiquity in Near Oceania and
Australia than in Southeast Asia (Allen, Gosden & White
1989: 550-551; Horton
1994:380-381). The external cognates are therefore regarded as
possible borrowings, and a higher-level reconstruction is not
proposed.
Chowning (1991: 55) argues that the gloss
‘earth oven’ is too restrictive, as raised ovens (covered with leaves
rather than earth) as well as pit ovens are common in Oceania.
The usual implement for handling hot stones and food in the oven was
tongs, probably a bent or folded piece of cane or midrib of a coconut
frond. Note that some terms have been derived from the transitive verb
form POc *kapit-i- ‘grasp (with tongs)’.
Lichtenberk adds a number of other terms which, as well as being used
in relation to oven cooking, probably also have more general
reference:
POc
*taqo(n), *taqon-i-
‘press down, weigh down with a weight, cover over with the cover
weighed down; close a stone oven, (when earth or the last layer of
leaves has been placed on top)’
‘take out the stones that have been heated in the earth oven;
stick used for same’
Only certain kinds of stones are suitable for use in stone ovens.
Although some terms for these have been located, e.g. Sudest (PT)
ɣərumu ‘cooking stones used for mumu’, Sursurunga (MM, New
Ireland) kor ‘kind of black stone used for mumuing’, no
cognates are identifiable.
Lichtenberk quotes from Oliver
(1989:49) who has this to say about
traditional cooking in Oceania: “Most cooking was by broiling, by
boiling (where there were clay pots, which was not everywhere), and by
baking (which was done in earth-pit ovens containing heated stones.)”
Lichtenberk continues (p.270): “In some places, food is boiled not over
fire but by means of hot stones dropped into a wooden bowl that contains
the food and some liquid. Food can also be steamed. In baking and
roasting, the food is often, though not necessarily, wrapped, usually in
leaves, sometimes in bark.”
He has listed terms for a range of cooking methods, with a word of
warning as to the reliability of precision in English glosses (‘roast’,
‘grill’, ‘broil’, ‘cook in fire’ etc.). He writes (p.273), “Even though
more than one term for roasting may be reconstructable for POc, the
differences in the glosses for the witnesses most likely reflect
differences in the English backgrounds of the authors rather than
differences in cooking methods.”
Lichtenberk also gives a reconstruction, *biti ‘cook with
hot stones (either by baking in stone oven, or by boiling by dropping
hot stones into liquid)’ (p.271), which he suggests may be POc, although
he points out that all cognates are from either Northwest or Southeast
Solomonic languages, and thus likely candidates for borrowing. We have
opted here to label his reconstruction PSS.
When food was roasted, instead of simply being placed in the fire, it
may have been skewered. The terms referring to skewering and skewers
were also used to refer to testing food by pricking to see whether it
was done. (See PMP *cukcuk, suksuk ‘skewer’, and POc
*(su)su(k) ‘anything used to pierce, prick, a skewer’;
*(su)suk-i- ‘pierce, prick’; Ch. 4, §3.2.1 and Ch. 9, §4.1).
Two verbs are reconstructable for POc with the meaning ‘boil’. The
problem is that the sources of data do not usually distinguish between
boiling and steaming, both of which are practised in Oceania, ‘boil’
being the usual gloss. The difference between boiling and steaming lies
in the amount of liquid used. For boiling, the food is (more or less)
entirely covered with the liquid; for steaming, only a small amount of
liquid is used. If POc did have a lexical distinction between the two
processes, there is some evidence, albeit weak, that the verb for ‘boil’
was *nasu and that the one for ‘steam’ was *napu, It
is also conceivable that of the two terms, *nasu was unmarked
and could be used to refer to both types of process.
Lichtenberk contributes a POc term which he glosses as ‘warm oneself
by fire; warm up, reheat (esp. food)’. Ann Chowning has added cognates
from Kove, Molima and Nakanai which, like the Pukapuka term, refer to
heating of leaves in preparation for mat making (see Chapter 4, §3.1.1).The POc gloss
has been generalised accordingly.
Two similar terms for wrapping food for cooking (baking or roasting),
POc *kapu(t), *kaput-i- and POc *kopu, have
been reconstructed. Mills (1981: 73) has proposed PMP
*kaput ‘close, cover’, while Blust (1986:45) gives a gloss of ‘tie or clasp
together’ for an identical PMP reconstruction. Oceanic reflexes indicate
two distinct terms, although there may have been some blending of
meaning in daughter languages. Lichtenberk suggests that the two POc
terms may have referred to different ways of wrapping food; Firth (1957), for instance, gives four terms
for different ways of wrapping food in Tikopia. From the glosses below,
it seems possible that *kapu(t), *kaput-i- referred to
covering (as an earth oven is covered), whilst *kopu referred
to the wrapping of the food in bundles. For both terms, reflexes
indicate wider reference than to food alone. In Polynesia they have come
to refer more specifically to clothing or blankets.
Lichtenberk also considers terms for uncooked food. He proposes a
general POc term *(a)mataq for raw food (possibly excluding
certain foodstuffs) which also refers to unripe foodstuff, such as
fruit. Its reflexes in fact seem to reflect two forms, *mataq
and *qa-mataq, the latter presumably derived from the former,
but the function of *qa- is unclear.
PAn/PMP *maqataq evidently consisted of *ma-
‘anticausative’ (Ch. 2,
§3.1.2) + qataq ‘eat something raw’. The POc reflex
*mataq shows irregular loss of medial *-q-. The
unaffixed base also survived in POc, and is reflected, with meanings
similar to *mataq, in some Oceanic languages.
Alongside *qataq, PAn also had a term *qetaq, which
retained its verbal use in POc, referring specifically to raw meat, fish
and shellfish, and to the eating of such foods:
Lichtenberk (p.277) records two main ways of preserving food in
Oceania. One is drying, either in the sun or above fire (the latter may
involve smoking the food); the other is fermentation.
Fermentation of food is geographically much more restricted. Small
islands, particularly atolls, are susceptible to food shortages as a
result of drought or prolonged stormy weather which makes fishing
impossible, and have greater need for food reserves. Food fermentation
is practised primarily in those areas where sago is not available as a
staple, the latter being easily stored for lengthy periods of time.
Assuming that sago was available as a foodstuff in POc times (see Dutton 1994), Lichtenberk argues
(p.278) that the nearly complementary distribution of sago as a staple
and fermented foodstuff suggests that the presence of fermentation in
Micronesia and Polynesia is due to independent developments. Also
relevant, however, is the fact that fermented breadfruit is regarded as
one of the few foods suitable for carrying on longer sea voyages, and
thus a prime suspect, along with its term, for borrowing. Yen (1975) writes that in Melanesia
fermentation appears to be practised only in Santa Cruz and ascribes its
presence there to Polynesian influence. However, David Walsh has pointed
out (pers.comm. with Lichtenberk) that fermentation is practised also in
some areas in Vanuatu. (See the Namakir form mada below.) Ross
(Food plants, vol. 3) points out that there is a curious cross-over
between the reflexes of PEOc *mara ‘preserved breadfruit’ and
POc *madraR ‘grow ripe, overripe’, *madraR-i- ‘grow
ripe, overripe from (s.t.) (?)’. In scattered languages in Vanuatu, Fiji
and Micronesia, a reflex of *madraR/*madraR-i- has
taken over the ‘preserved breadfruit’ sense which we would instead
expect to find associated with a reflex *mara. Since closely
related languages both in Vanuatu and in Micronesia disagree as to which
term they reflect, Ross infers that the closeness of both form and
meaning has led to confusion between the two terms, leading to more than
one independent occurrence of crossover. POc *madraR seems to
have been applied specifically to breadfruit and bananas, both of which
become soft and mushy when overripe. Breadfruit and bananas are also the
two foodstuffs which today are most commonly preserved by fermentation
(Yen 1975). Chowning lists a Nakanai
cognate which indicates that the original POc term may have meant simply
‘be spoiled (of food or drink)’, with its meaning narrowing in Eastern
Oceanic.
We have terms for the action of pounding or crushing, but have been
unable to reconstruct any term for the implement or implements involved
apart from PWOc *walu ‘sago beater’ and POc *ike
‘(bark-cloth) beater’ (for the latter, see Ch. 4, §S . l).
Sago is pounded with an implement consisting of a head like a small
adze, mounted on a long handle (Ann Chowning pers.comm.). Smaller wooden
pestles are used for mashing foodstuffs like taro, banana and
breadfruit, and for crushing nuts, particularly galip
(Canarium) and betel (Areca) nuts. Sometimes stones are used,
or wooden pestles with stone heads. Lithgow lists Muyuw (PT)
kilakil ‘stone-headed sago pounder’, and Chowning (pers.comm.)
records a Nakanai (MM) term mulumu ‘pestle of wood or bone used
for crushing almonds or areca nuts’, from the verb mumu ‘crush
in a mortar’. Lichtenberk write (p.277) that:
Pounded foods are widespread in Oceania (Yen 1975). In Pidgin and in English they
are referred to as ‘puddings’ (although not all puddings require
pounding). The usual main ingredient is tubers (taro, yam), cooked and
pounded, mashed into a paste together with other ingredients such as
coconut cream or nuts, and then usually recooked.
A large wooden bowl (POc *tabiRa, PCP *kumete, see
Ch. 4, §2.2.1) was
traditionally used as a mortar. Sago was pounded in a trough, stronger
and heavier than a bowl, and designed so that water could run off from
one end. Only isolated terms have been located (Muyuw kas
‘trough for sago-making’; Loniu kupʷi ‘trough in which sago is
pounded’).
As Blust (ACD) notes, terms dealing with the action of pounding
frequently contain the root *-tuk. A number of reconstructions
have been made which evidently refer to hitting, pounding, beating,
breaking open, and so on, not just of foodstuffs, but with general
application. Cognate sets for the following reconstructions are included
in Chapter 9, §5.1: POc
*tutuk ‘pound, mash by pounding, hammer, crack by hammering’
(from PAn *tuqtuq); POc *tuki- ‘pound’; POc
*putu(k) ‘knock, pound, beat’ (ACD) (from PMP *buTuk);
and POc *qatuŋ-i ‘strike from above, pound’.
Foodstuffs such as tubers and coconut meat are often grated before
further processing. Tubers also need to be scraped to remove dirt. Both
processes can be carried out with a shell such as a cockle, and the
scraper/grater may be referred to by its shell name. Scrapers and
graters are not always separately identified. The same implement may
serve both functions, just as the same verb may refer to both
actions.
‘bivalve mollusc (Asaphis violascens), possibly other
related bivalves also; shell traditionally used as cutting or scraping
implement, as food scraper for coconut, breadfruit’
PEOc
*kaRi
‘scraper; bivalve sp., used as a scraper’
(Geraghty 1990)
A number of formally similar terms have been reconstructed for the
action of scratching, scraping and grating. Although the reference is
often primarily to preparation of foodstuffs, such terms are often used
to refer to any similar action such as scratching the skin or scraping
the bottom of a canoe.
Yet another reconstruction is POc *asa(q) ‘grate, sharpen by
grating or rubbing’ from PAn *Sasaq ‘whet, sharpen’ (see Ch. 4, §4.1.S).
Lichtenberk comments (pp. 280-281) on the recurring sound symbolism
evident in these items, a feature which Blust has identified with PAn
forms meaning ‘rub, scrape, scratch’. Terms with similar sound symbolism
readily undergo conflation. Lichtenberk also refers to the possibility
that in POc or pre-POc times there was some kind of derivational
relation among at least some of the forms, noting that unlike the other
forms, *ko(r,R)as took as its direct object a noun phrase
referring to the stuff scraped off rather than the object being scraped
(e.g. the coconut meat rather than the coconut). (See also Ch. 9, §2.2.)
Lichtenberk lists a number of terms for the action of peeling. The
first is a transitive verb formed from the base *kulit ‘skin’.
See also POc *supi ‘sharpen, shave, pare’ (Ch. 9, §3.6) some of whose
reflexes also mean ‘peel’.
PAn
*kulit
[N] ‘skin’; [V] ‘peel, remove skin of s.t.’
POc
*kulit, *kulit-i-
[N] ‘skin’; [v] ‘peel, remove skin of s.t., bark (a
tree)’
The following term may have referred both to the husking of coconuts
and to the implement used. This was a sharpened stick, usually set
finnly in the ground. The same term is sometimes used to refer to a
digging stick.
POc *soka, soka-i- (V) ‘pierce; stab’ (see Ch. 9, §4.1) (Lichtenberk has
*joka) and POc *potak ’crack or split open (nuts,
coconuts) are also used of husking coconuts:
The most common method of straining, necessary for processing such
foodstuffs as sago and coconut cream, and also used in the production of
kava, is by using the fibrous spathe of a coconut frond.
Lichtenberk (1994) reconstructed PEOc
*unu(p) ‘strainer, probably the fibrous spathe of a coconut
frond’, but it is reasonably certain that the reflexes attributed
thereto should be combined with those above and attributed to
*Runut/*nuRut (Ross
1996d). Lichtenberk reconstructed final *-p on the basis of
Mota (NCV) unuv ‘(a fluid) sink in, be absorbed’, but we take
this not to be cognate. As Geraghty (1990) points out, Sa’a (SES) shows
irregular loss of initial *R-, but Rennellese (pn)
gunu, borrowed from a SES language, reflects Proto SES
*lunu (< POc *Runut). Sa’a also shows loss of
*R in a few other items, so its loss here is not completely
surprising.
Another reconstruction, POc *piri ‘twist, wrap around’,
whose reflexes refer typically to the manufacture of rope or twine (see
Ch. 4, §3.2), has reflexes in
some languages which refer to extracting coconut milk by squeezing it
through coconut fibre (MM: Kia piriki ‘wring, squeeze coconut
milk’, NCV: Tamambo viri ‘twist, plait, braid; coconut milk’.
See also SES: Arosi biri-yi ‘squeeze coconut milk through
fibre’).
Lichtenberk writes (p.283):
It is not possible to detennine fully what the distinctions were
among the various terms for extracting liquid, but a number of contrasts
can be postulated. Items *pipik, *pisak,
*poji, *losi took as their direct object a noun phrase
referring to the object out of which liquid is extracted, whereas
*poRos and *pirik took as their direct object a noun
phrase referring to the liquid extracted. *pipik and
*pisak referred to squeezing by pressing, while *poji
and *losi referred to squeezing by wringing. The item
*pirik referred specifically to a wringing action, while
*poRos may have been an unmarked term whose meaning subsumed
both pressing and wringing. (As Ann Chowning has pointed out
[pers.comm.] there are two basic ways used to extract coconut cream:
wringing coconut gratings through coconut ‘cloth’, or squeezing them in
one’s hands.)
Allen, Gosden & White 1989 “Human pleistocene adaptations in the tropical Island Pacific: recent evidence from New Ireland, a greater Australian outlier”
