4.14 Types of reduplication.

In Chapter 3 we discussed Initial Reduplication (rule P-2) and Final Reduplication (P-3). Only P-2 is relevant to nouns. As explained in §3.3, the basic generalisation is that monosyllables and stems beginning with a stop have monosyllabic reduplication /C1V1-/ while polysyllabic stems beginning with something other than a stop have a bisyllabic reduplication /C1V1C2V2/ Further complications are discussed in that section.

For nouns, the regular grammatical function of reduplication by P-2 is (3+) plural marking. This function is found with a large number of adjectival nouns (NAdj), and hence with translation equivalents of many English human nouns (other than kin terms and Personal names). However, many such nouns take a Pl prefix such as /mij-/ instead of reduplicating, so the Pl must normally be specified in dictionary entries. Reduplicative NAdj may be used for nonhuman Pl reference. However, (nonadjectival) nonhuman nouns themselves have no plural marking (except for collectives, see below).

Some exx. of regular Pl reduplication by P-2 for NAdj with plural (more than three) sense are these:

SgPlgloss
badirinyaba-badirinyaghost
dhudabadadhu-dhudabadawhite man
lhalmarlhalma-lhalmarforeigner
ngajiwanggungaja-ngajiwangguold woman
(In nonpredicative use, the Sg forms shown would be modified by adding the relevant HumSg suffix, except that this is only optional for ’ghost'.)

There are also some high-frequency irregular plurals which in some cases look like irregular reduplications:

SgPlgloss
manungmananungwoman
w2irigw2ura:yungsmall
w2inyigw2unya:nyungsmall

(The same remark about use of HumSg suffix made above applies to these items as well.) The two Pl forms for 'small' are also unusual in that, with human reference, Pl prefix /mij-/ is added, hence /mij-bura:yung/ and /mij-bunya:nyung/ 'children'.

The two stems for 'small' consistently use the irregular Pl form for plural sense with nonhuman as well as human nouns. Since nonhuman nouns (except for a handful of collectives, see below) do not mark plural, either in their own affixes or in cross-referencing pronominals in the verb, the Pl form of a modifying adjective like 'small' may be the only evidence of (cryptotypic) plurality of such nouns. For example, /mada/ 'grass' has no direct Pl form, and since the English counterpart is usually a mass noun we might assume this is also true of the Nunggubuyu word. However, /mada/ may co-occur with a NAdj like 'small', and in this case we may get /w2unya:nyung/ suggesting a (usually hidden) underlying plurality.

Some other NAdj with similar high-frequency Pl forms used with nonhuman as well as human referent are these:

SgPlgloss
lhamungurlhama-lhamungurshort
ala:diala:-'la:dibad
runggalrunggu-runggalbig

On the other hand, for many of the less common NAdj which take a reduplicative Pl, the reduplication is sometimes or even frequently omitted. (For //lh///dh/ in 'short', see P-19.)

In addition to the Pl reduplication just illustrated, we have at least two other types of nominal reduplication using essentially the same phonological rule (P-2): collective and adjectival. The term 'collective' is applied to a few topographic nouns which reduplicate to indicate a substantial but indefinite number of instances:

SgCollectivegloss
alaala:-'lariver
wuruguwuru-wurugubillabong
riljirilji-riljiisland

The adjectival type involves reduplication along with a suffix /-j/ which may ablaut the preceding vowel to /u/ The reduplication is obligatory in most or all of the stems I have in mind, but there is often a cognate verb which may fluctuate between simple and reduplicated form. In any event, the reduplication is not specifically plural or collective in the noun forms (and need not be specifically repetitive or prolonged in the verbs). See §14.15.

VerbNAdj
=w2ara-'be crooked'w2ara-waru-j'crooked' (cf. P-19)
=w2ilwila-'be pliant'w2il-wilwilu-j'pliant' ( " " )
=duma-'be black'du-duma-j'black'
=ngalngala-'be white'ngal-ngalngalu-j'white'

There are also a number of frozen formations of this type for which a synchronic segmentation is questionable, as with /jarmayarmaj/ 'long, tall'. There are also some cases where a noun or NAdj is derived from a simple noun root by reduplication and addition of /-yij/ as in /mada/ 'grass', derivative /mada-mada-yij/ 'having lots of grass'. For more details and exx. see §14.14. It should be noted that in the exx. 'crooked' and 'pliant' just given, the //w2// becomes /b/ by Leftward-Hardening P-19 in some contexts, just as //lh// in 'short' (above, this page) becomes /dh/.

In addition to plural, collective, and adjectival reduplications, there are a few other possibly distinct minor types. The time-of-day adverbial nouns /ngamugijgaj/ 'morning' and /adharwara/ 'dusk' are often found in reduplicated form with no specifically repetitive or plural sense: /ngama-ngamugijgaj/ /adha:-'dharwara/ This may be related to the tendency of cardinal-direction adverbs to take special reduplicative forms (§7.29).

From /mu:n/ 'foot' a form /mu-mu(:)n-gala/ 'on foot' is recorded (NMET 163.4). The suffix is Ablative /-w1ala/ but it looks as though a special reduplicative form is used here.

Alongside adverbial noun /ragij/ 'first (in place)' we have a more or less equivalent form /raga-ragi-j/ which has no specific Pl meaning. Perhaps this is related to the isolated reduplications of other adverbial forms just mentioned.

From /rangag/ 'tree' (ANAwu class) or 'dugout canoe' (MANA), we have /ranga-rangag/ 'wooden seats in canoe' (ANAwu). Here the reduplication has an idiosyncratic derivational function. The various nominal reduplications we have seen occur at the beginning of the root or stem, and do not include the NCinfl prefix. Even with monosyllabic roots, suffixes such as case markers cannot be copied in the reduplicative segment. NCder prefixes are part of the stem and thus are affected by reduplication, so that //Rdp-uG-muwaj// shows up as //uGmu-uG-muwaj///(w)umu:-'-muwaj/ 'named (human Pl)', with //-uG-// as Pl NCder prefix.