Kaluli

Timothy Usher, Santa Fe Institute

Kaluli, also known as Bosavi, is spoken by approximately 2,500 people (2004) living in villages around 750-800 meters above sea level on the northern and western slopes of Mount Bosavi, between the Libano and Rentoul rivers to the north and west and the Bifo river to the east, in Papua New Guinea's Southern Highlands Province just east of the border with Western Province. Speakers call themselves Kaluli /kʰaɺu-ɺi/, meaning “real people,” and consider themselves part of a larger group Bosabi Kalu /bosabi kʰaɺu/ which includes speakers of at least the Kasua, Sonia and Onobasulu languages (Shaw 1973: 195, Voorhoeve 1975: 394, Schieffelin and Feld 1998: xi, Grosh and Grosh 2000: 4, 2004a: 7, 2004b: 59, Bodani 2006: 1, q.v. Schieffelin and Feld 1998: ix, 22.)

The Kaluli divide their language into four dialects, Kaluli, Ologo in the east and Walulu and Kugenesi (Wisesi) in the west, as follows; however Grosh and Grosh were unable to find any linguistic correlaries to this division (Schieffelin and Feld 1998: xi, Grosh and Grosh 2000: 1, 4-5):

Kaluli

Ologo – Wasu, Didesa

Kaluli – Wabamisen, Wogodagan, Suguniga, Bona, Muluma

Walulu – Orabia, Tabili, Wanageso, Gambano, Sibalama, Wasoweido, Baniso

Kugenesi – no current village, living in Wanageso and other villages

Baniso village is properly Sonia-speaking, but the Sonia language is being replaced by Walulu, which has in turn adopted some terms from Sonia.

Rule (1964) phonology and grammar of Kaluli (unobtained)

Anonymous and Rule (2009) sketch phonology of Kaluli

McElhanon and Voorhoeve (1970) include Kaluli comparisons under the name Bosavi

Franklin (ed. 1973: 558) 95 comparative terms for Kaluli

Shaw (1986: 66) 99 comparative terms for Kaluli

Schieffelin and Feld (1998) dictionary of Bosavi

Rumsey San Roque and Schieffelin (2012) Kaluli ergative

Grosh and Grosh (1991) sociolinguistic survey of Kaluli (unobtained)

Grosh and Grosh (1993) sociolinguistic survey of Kaluli (unobtained)

Grosh and Grosh (2000) sociolinguistic survey of Kaluli

Grosh and Grosh (2004) grammar of Kaluli of Suguniga and Didesa villages

Grosh and Grosh (2004) phonology of Kaluli

Schieffelin and Feld (1998: xiv-xvi, cf. Grosh and Grosh 2004b: 60, ibid., Anonymous and Rule 2009) give 13 consonants and 7 vowels for Kaluli as follows:

Vowels are further distinguished by the presence or absence of nasalization:

Grosh and Grosh write what are given here as voiced stops /b d g/ as plain voiceless <p t k>. Anonymous and Rule (2009) assert that voiced /b d g/ are realized as voiceless [p t k] initially, but Grosh and Grosh (pp. 60-62,) give a number of examples in which initial plain voiceless stops are realized as voiced, and (p. 64) state that medials are usually but not always voiced, while slight voicing may be heard initially. Schieffen and Feld (pp. xiv-xv) argue the opposite, stating that /b d/ are always realized as devoiced [p t] despite Rule's characterization, although they qualify that medial velar /g/ is lenited to a fricative [ɣ] adjacent to vowels /o ɔ a/. We represent this series as voiced as <b d g> in accordance with Rule's and Schieffelin and Feld's orthography as this is what they are both systemically and historically.

Voiceless aspirated stops /tʰ kʰ/ do not occur finally, and occur medially only in compounds and loans.

Fricative /f/ is specified as labiodental.

Apical non-stop /ɺ/ is specified as a flapped lateral. When followed by rounded back vowels /u o ɔ/ it is realized as retroflexed [ɭ]. It does not occur initially.

Voiceless aspirated stops /tʰ kʰ/ may be realized as fricatives [s x], with some speakers merging /tʰ/ with [s] consistently and all those who realize /kʰ/​ as [x] doing so consistently (Schieffelin and Feld 1998: xv, xvi, Grosh and Grosh 2004b: 64) As Papuan Plateau voiceless aspirated bilabial /*pʰ/ has already been fricated to /f/, these trends may be viewed as continuing this process, the end result of which is to eliminate contrastive voicing altogether as follows, bearing in mind that initial voiced stops are realized as voiceless unaspirated stops initially, thus voiced medial stops would be considered allophones of these:

This is very similar to the systems of the Mount Sisa languages to the northeast and of Kasua to the southwest, except that the former have merged initial voiceless apical stop /*t/ with /d/ rather than with /s/, while both merge velar stops /*k *g/ rather than fricating /*k/ to /x/. The resulting distinction between velar and unoccluded fricatives /x h/ is typologically unusual from a New Guinean perspective. The frication of /*t/ to /s/ is shared with Onabasulu to the north as well as with the closely-related Kasua language to the southeast.

According to Schieffelin and Feld (p. xiv,) some speakers realize low mid front vowel /ɛ/ as low front [æ], to which it regularly corresponds in Kasua.

Only a restricted subset of consonants occurs finally:

Final nasals and non-stop /m n ɺ/ are realized as devoiced [m̥ n̥ ɺ̥].

Final voiced stops /b g/ are realized as devoiced and unreleased [p̚ k̚].

Final phonetic glides [w j] are interpreted as sequential high vowels /u i/ in a phonetic diphthong (below.)

Consonant clusters do not occur synchronically, although the previous existence of clusters across morpheme boundaries is evident in the inflections third-person possessed kin terms and consonant-final verb roots (below.)

Vowel sequences are of two types. Where the second vowel in the sequence is of equal or greater height than the first, the sequence is heard as a diphthong and constitutes a single syllable [eⁱ oⁱ ɛⁱ ɔⁱ aⁱ ɛᵘ ɔᵘ aᵘ aᵒ]. Where the second vowel is lower than the first, each vowel is its own syllable, excepting only that sequences /kʰuV guV/ are realized as [kʰʷV gʷV]. Sequences of two identical vowels do not generally occur. Observed vowel sequences include:

Schieffelin and Feld (1998: xvi, Grosh and Grosh 2004b: 63) are somewhat equivocal in their characterization of the distribution of nasalized vowels, stating that nasalization is found on only a small number of words where it is irreducible, but that there are no known minimal pairs. Where a nasalized vowel is followed by a voiced stop /b d g/, the stop is realized as prenasalized [mb nd ŋg].

Rule proposes a system of three pitch tone levels for Kaluli syllables in which high tone and low tone are contrastive. Schieffelin and Feld (1998: xvi-xviii) reanalyze this as stress. Most often, monosyllables are unstressed, disyllables are stressed on the first syllable, and trisyllables are stressed on the first syllable except for first person future forms which take penultimate stress, as do words with four syllables. Stress is changed, generally to the penult, by the addition of affixes. Schieffelin and Feld indicate stress only where it is contrastive; Grosh and Grosh don't indicate it at all.

Grosh and Grosh (2004a: 35-42) give pronouns in eleven case forms for Kaluli. We divide these into several groups according to their shared bases as follows:

The singular forms of stems I and III are identical to the possessive proclitics and prefixes found on class II and class I kin terms respectively (below.)

Upon these bases are appended suffixes deriving case forms as follows:

The bases of the unmarked case used for both subject and object are designated as stem I. These are also the bases of the dative, used to indicate recipients of items and actions, and the limiter case, used to contrast a specific reference to some other entity:

Pronouns in the unmarked possessive case precede the nominal they modify, and excepting the third person singular are identical to the stem II bases. The focused subject/object case highlights a focused referent in a clause, most typically but not limited to ergative referents. The possessive complement case is used to refer to something which is possessed (“mine” etc.) Relative pronouns are used in nominalized clauses. It is not really clear that the contrastive pronouns, which are used to emphasize that the agent is not who would be expected from a preceding clause, are based upon stem II forms, but there is nothing either to contradict it so they are presented here:

The possessive exclusive highlights the possessor in contrast to others. The singulars and plurals of the reflexive case, used when the agent of an action is also its object, are based upon those of the possessive exclusive. Dual possessive exclusives are not attested; the second and third person dual reflexives appear to be suffixed stem II forms (above):

Exclusive pronouns are used to highlight a participant to the exclusion of others. The suffixes used to indicate this case in the singular, plural and dual respectively are completely different:

In addition to the free pronouns given above, Grosh and Grosh (2004a: 15) provide a list of 20 kin terms, most of which which take one of three prefixed or procliticized inalienable possessors differentiated by person but not by number. These are identical to stem III and stem I respectively of the free pronouns (above,) except for the third person prefix which is unique to the possessive construction (though may be found on the stem III form of the third person dual /idɛni(n)/)::

Suppletive forms aside, inalienably possessed kin terms may be divided according to their patterns of prefixation into classes with subclasses as follows:

Class I roots take prefixed possessors. Probably the oldest layer of these is represented by meanings for which the first and second persons are indicated by prefixes /n- g-/ and the third person by /in-/. When prefixed first and second singular possessors are followed by consonant initial roots, a copy vowel of the first vowel of the root, or a high vowel if the root begins with a glide, is interpolated between them. Final /n/ in the third person disappears in some contexts and alters the following consonant in others, a phenomenon also found in the inflections of verb roots with final /n/ (below):

In one example, the prefixes are regular but the root of the third person form is suppletive; cf. /sɔwa/ “son”:

(Grosh and Grosh give /inso/; cf. Schieffelin and Feld 1998: 131)

One example has an unprefixed suppletive form in the first person and a (probably) prefixed suppletive form in the third:

In the second subclass, the third person possessive proclitic is found before a prefix which is superficially identical to that of the first person, but more likely reflects third person /in-/:

In the third subclass, the third person possessor is signified (or not) by zero:

One example has an unprefixed suppletive form in the third person:

In one example, the independent third person singular possessive form is given:

Class II roots take proclitic possessors in the second and third persons, but usually zero in the first, a fact which might be interpreted to suggest that they originated as address forms:

In a smaller number of examples, the first person is likewise signified by a proclitic:

Kaluli has a body-part counting system of the type characteristic of the New Guinea region in which the term for the number is identical to that for the body part which is touched during tallying. Counting begins from the left pinkie at one, proceeding to the thumb and then up the left side of the body until reaching the tip of the nose, then proceeding downward on the right side of the body as follows (Schieffelin and Feld 1998: 173-174, Grosh and Grosh 2004a: 45-57):

[under construction]

… (Grosh and Grosh 2004a: 50-52):

[under construction]

Portions of the presentation which follows are based upon Grosh and Grosh's (2004a: 16-30) description of Kaluli verbal morphology, but with significant modifications.

Suffix-initial consonants and vowels which are dropped when following consonant-final and vowel-final roots respectively are indicated with parentheses below. In a number of Grosh and Grosh's representations this consonant was not indicated; we have added them here after their examples and those of Schieffelin and Feld (1998.) This is the first important difference between our analyses as it allows us to reduce most stem variations to the regular application of phonological rules (below.)

We depart from Grosh and Grosh's (2004a: 16, 16-26) terminology in designating their STEM2 and STEM3 as stem III and stem II respectively, thus altering Schieffelin and Feld's (1998: xi-xii, ibid.) order of presentation. This was done because stems I and II share phonological similarities vis-a-vis stems III and IV, with stem I being maintained as such because it, followed by the imperative suffix, is Schieffelin and Feld's citation form, as well as that used when followed by another root in a compound.

Person is distinguished in only two tenses. These are likely the remnant of a once-broader pattern:

(Grosh and Grosh give /-(m)eib/ second and third person future while examples in Schieffelin and Feld give /-(m)ɛib/; we follow the latter.)

Final (i.e. independent) verbs take tense/aspect/mood suffixes as follows:

Medial (i.e. dependent) verbs take tense/aspect/mood suffixes as follows:

It may be seen that all of these suffixes underlyingly begin with bilabial nasal /m/ or apical non-stop /ɺ/, excepting only the past which is a single vowel. Their phonological shape along with those of the roots to which they are appended (below) determine the realizations of the stem and suffix alike.

The reason for the difference between stem I forms and those which take stem II is not immediately clear, as both are followed by suffixes beginning with underlying /m/. The answer would seem to be stress, since imperatives are monosyllabic but stem II forms are not; indeed Schieffelin and Feld (1998: xviii) state explicitly that the addition of first person future /-(m)ɛnɔ/ shifts stress to the penult, which the addition of (e.g.) imperative /-ma/ does not. This would explain all of Grash and Grosh's forms excepting the different subject purposive /-(m)ɛki/, which their examples consistently show to take stem I forms (/maj-ækʰi/ “eat”, /di-mækʰi/ “take”, /asuw-ɛ-fo-mækʰi/ “help”.) This in turn might be solved by observing that the segment /kʰi/ must itself represent a suffix, as is mandated on phonological grounds and supported by its frequency as a final element in the forms above; since the other suffixes with initial /m/ upon which /-kʰi/ appears were multisyllabic already so would be assigned to stem II.

Grosh and Grosh give three aspect auxilliaries as follows:

According to Grosh and Grosh, the vowel of the future continuative harmonizes with that of the following future suffix, which is first person /-ɛnɔ/ or second and third person /-eib/, thus either /-mɛɺ-ɛnɔ/ or /-meɺ-eib/; however as noted above Schieffelin and Feld give the latter suffix as /-ɛib/, which we credit in part because the vowel sequence /ei/ is otherwise only very marginally attested (e.g. tʰuɺunei/ “ground dove”) while Grosh and Grosh's own phonology (2004b: 67) gives this suffix as /-ɛib/. Moreover, it's clear that both /-meɺ-/ and Grosh and Grosh's present continuative /-ɛɺ-/ are respectively the stem Ii and stem III forms of /meɺe-/ “stand; wait, exist, stay” in compounds with the preceding stems (below;) it is not the case that this occurs only with future tense suffixes as is implicitly suggested in Grosh and Grosh. Schieffelin and Feld give the variation /-mɛɺ-ɛnɔ/ in only one example, casting doubt upon the general validity of this phenomenon.

More doubtful still is the claim that the suffixation of /meɺ-/ induces the stem III form of the preceding root, as this would contradict the phonological bases governing stem selection as well as compounding rules which generally require stem I in the first member (below;) in fact the results would be nearly incoherent as stem III forms are specifically the results of syllable reduction (or not) or clustering with following underlying suffix-initial /ɺ/. For instance the verb /do-fo-/, one of two examples given (p. 26,) would become in stem III merely /do-ø-/ while strengthening following /ɺ/ to [tʰ s] (see auxilliary /-fo-/ below) as seen in Schieffelin and Feld (1998: 41.)

In addition to these, there are six suffixes which follow inflected final verb forms, all but one of which mark the evidentiary status of an action. The uncertainty suffix must be viewed as an enclitic, as it is found also on nominals:

… prefixes … (Grosh and Grosh 2004a: 52, 58-60)

Kaluli verbs can be divided into classes according to the pattern of variation between their stem forms, which is determined by the phonotactic shape and the final segment(s) of the underlying root. The most fundamental divisions are those between multisyllabic roots and monosyllabic roots, which are uncommon but highly salient in meaning, and between roots with final vowels and those with one of a restricted set of final consonants.

For a number of contrasts in the realizations of multisyllabic verb roots, these final segments are not alone sufficient to explain the variation. In these instances we have posited contrastive stress <ˈ>, possibly at the pre-Kaluli stage; we do not assert that this stress is synchronically present. It is not clear if this is the same phenomenon as Rule's tone, reanalyzed by Schieffelin and Feld (1998: xvi-xviii) as stress, which is shown here as <V́> as in Schieffelin and Feld's dictionary. Accordinly, we refer to our as stress and to Schieffelin and Feld's <V́> as tone.

Verb roots are of the following shapes:

It has been noted above that apical voiceless aspirated stop /tʰ/ may be realized as fricative [s], with some speakers of the central (Kaluli) dialect merging these consistently (Schieffelin and Feld 1998: xv, xvi, Grosh and Grosh 2004b: 64.) As Schieffelin and Feld's dictionary is based promarily upon the central dialect, most of these entries are ordered as <s> with variations indicated in the entry. Only a small minority of forms are given with invariable [tʰ] (q.v. pp. 138-140.) To simplify the presentation which follows, sounds which vary between [s tʰ] are given here simply as <tʰ>.

Patterns of suffixation are summarized as follows, with present imperative /-ma/, future imperative /-(ɺ)ɛbi/, first person future /-(m)ɛnɔ/ and past /-ɔ/ suffixes (above) illustrating the dispositions of stems I through IV as found in Schieffelin and Feld (1998) (though with stems II and III swapped as noted above):

No verbs have been found with stem final /mV-/, while stem-final /m-/ has been found only in the highly abberant monosyllable /ˈham-/ “go” (below.) Presumably these have become /nV-/, as occurs intermittently with now-final /m/ in nominals, or has disappeared, perhaps due to the strong aversion to consecutive occurences of /m/ as seen in several imperative forms (below) in conjunction with the ubiquity of stem I and stem II suffixes (above) which begin with underlying /-m/.

These patterns of stem variation are exemplified in the charts which follow. Where verb paradigms are defective, for example giving a form with an auxilliary in some tenses but not others, or otherwise suppletive, the gaps in the resulting split paradigm are indicated with dashes. Since monosyllabic verb roots are less common and undergo more surprising changes, their paradigms are presented after those of mutlisyllables. As Schieffelin and Feld's (1998) glosses, especially for compounds, are often too involved to recapitulate here, it is recommended that one consult the dictionary to gain a better sense of their meanings.

We begin with multisyllabic roots with final vowels, because these are the most straightfoward of Kaluli's verbal paradigms.

…

…

… /nV/ … :

… :

Two examples which likewise have stem-final [nj] rather than [j] in stem III and stem IV forms, /gɔɺɔne-/ and /hamane-/, are presented as adjectives (below) and are thus exemplified only by stem III and stem IV forms with second/third person present and past suffixes respectively; their stem I forms are known from their occurence in compounds (below.)

Disyllables of the form /Cona-/ become /Cu(w)-/ in stem III and stem IV forms:

… /dV-/ …:

…

Two examples unexpectedly have /ɺ/ rather than /d/ in the past tense and future imperative respectively:

… multisyllabic roots where unstressed stem-final /di- de-/ follow low mid and low vowels /ɛ a ɔ/ in the preceding syllable, or where /du-/ follows /ɔ/ … the lack of examples in which low mid front /ɛ/ or low central /a/ in followed by /u/ the final syllable is likely due to the fact that these would have already been alterered to low mid back rounded /ɔ/ … in stem III and stem IV forms, where [d] is ratained rather than leniting to [ɺ], the high vowel is metathesized to form a diphthong /ɛi ɔi ai ɔu/ in the preceding syllable prior to its loss. Because vowel sequences /ɛe ɔe ae/ do not occur, metathesized /e/ is raised to /i/

…

…

…

… stress … /ˈdV-/ …:

One example is ambiguous between stressed root-final /ˈdV-/ and unstressed /dV-/ (above,) as the only difference between many of these forms, excepting those with unstrssed /dV-/ that undergo vowel metathesis, is the retention of /d/ vs. lenition to non-stop /ɺ/ in stem I and stem II forms:

…

…:

…

… stress …:

One example shows an irregular reduction in the stem III form, possibly because, unlike the preceding examples, the base is trusyllabic. This is distinct from the reflex of unstressed stem final /gV-/ which would be /duɺug-ɛbi/ (above):

… …

… …

…

One example unexpectedly takes suffix /-ɛ/ in the past:

… stress … All known examples of this pattern have final high vowel /i u/ or low mid front /ɛ/ …:

…

The most common pattern is that of multisyllabic roots with final segment /ɺV-/ and penultimate stress. In the future imperative underlying /ɺV-ɺɛbi/ is collapsed to /ɺ-ɛbi/ while in the past tense the final vowel is dropped and replaced by /-ɔ/:

For one of these roots, an alternative past form is given with /u/ retained /-ɔ/ dropped in the past; however the loss of final /u/ in the stem III form shows penultimate stress to be original:

… final stress … As with stressed root-final /ˈsV-/ (above,) all known examples of this pattern have final high vowel /i u/ or low mid front /ɛ/, suggesting that only these vowels can displace that of the past /-ɔ/. Roots with stressed final segment /ɺɛ-/ are distinguished from those with auxilliary /-dɛ-/ [-ɺɛ-] (below) in the stem III form, which remains as /ɺɛ-ɺɛbi/ rather than collapsing to [ɺ-ɛbi]:

For one of these roots, an alternative past form is given with /-ɔ/; however the rentention of final /i/ in the stem III form shows final stress to be original:

If the vowel of a stem-final stressed syllable /ˈɺV-/ is low central /a/, it is dropped in the past while past /-ɔ/ is retained. These are still distinguishable from unstressed stem-final /ɺa-/ (above) because in the future imperative underlying /ɺa-ɺɛbi/ is not collapsed to [ɺ-ɛbi]:

… multisyllabic roots with final consonants … We assume stress, which is not contrastive in these paradigms, to fall upon the final syllable of the root, because final consonants in nominals invariably result from the loss of destressed final vowels and because Grosh and Grosh (2005b: 69) state that words with final consonants tend to take final stress.

…. final apical nasal /n/ …:

…:

…:

…

Roots with final apical voiced stop /d-/, like those with final nasal /-n/ above, lose final /d/ in stem I and stem II forms, but it surfaces in stem III and stem IV:

Roots with final apical non-stop /ɺ/ follow the same pattern as those with final voiced stop /d-/ above, with final /ɺ/being lost in stem I and stem II forms but surfacing in stem III and stem IV:

Two apparently irregular examples are ambiguous between final /ɺ-/ and unstressed final /ɺe-/ (above); the first of these is also found in a compound /nana-tʰanˈden-/ (below) with the stem I form /nana-/:

Monosyllabic roots with final consonants …

… monosyllabic roots with final bilabial nasal /m/ …:

Monosyllabic roots with final bilabial voiced stop /b/ …:

The root /bɔb-/ is also found in compounds with the auxilliary /-fo-/ (below) followed by the irregular verbs /ˈmese-/ “sit” and /ˈmeɺe-/ “stand” (below,) in which it is realized as /bɔ-fɔ-/.

Monosyllabic roots with final velar voiced stop /g/ …:

…:

…:

Monosyllabic roots with final labiodental fricative /f/ …:

…

…

Here Schieffelin and Feld's apical voiced stop /d/ rather than voiceless /tʰ/ in the third person future imperative (stem III) suffix is unexpected and unexplained:

… monosyllabic roots with final apical nasal /n/ …:

…:

… monosyllabic root with final fricative /s/. …:

Monosyllabic roots with final apical non-stop /ɺ/ following vowels /i e o u/ differ from roots of the form /ˈCVɺV-/ (above) only in the stem II form, in which /ɺ/ is dropped:

Monosyllabic roots in which final apical non-stop /ɺ/ follows low central vowel /a/ drop /ɺ/ in both stem I and stem II forms:

…:

…:

Monosyllabic roots with final vowels generally retain this vowel in all conjugations, excepting only those with final low central /a/ (below.) Generally this entails the loss of past tense /-ɔ/, although there are a few exceptions as will be shown.. Thus, we can conclude that all monosyllabic roots are stressed.

…

The paradigm for “give” is identical to these except that imperative /-ma/ has regularly become /-na/ due to the influence of root-initial /m-/ (below):

One apparent exception perhaps suggests the underlying form /hiˈji-/:

… As with “give” above, initial /m/ provokes a change in the infinitive suffix from /-ma/ to /-na/:

…

Monosyllabic roots with final vowel /a/ interpolate a subphonemic palatal glide [j] between the root and past /-ɔ/, where /Ca/ or /Cɔ/ might otherwise be expected. This is analogous to the realizations of ergative, locative and topic enclitics on nominals which likewise take an epemthetic [j] when following root-final /a/ (above,) except that here the vowel of the verb root becomes either high front /i/ or high mid front /e/.

…:

…

…

…:

…

On the strength of Grosh and Grosh's (2004b: 66-67) assertion that vowel sequences in which the first vowel is higher than the second are realized as phonetic diphthongs, two examples involving sequences /ɛu au/ are treated as monosyllabic verbs with a final vowel [ɛᵘ aᵘ].

…

…

Grosh and Grosh (2004b: 71, cf. 2004a: 10) observe that “Some verb stems tend to display vowel harmony in which the first vowel int he verbal suffix influences the preceding vowel(s) in the verb root. This vowel harmony process occurs only in a restricted set of verbs stems, and appears to be restricted to the domain of verbs with single syllable stems.” Here we outline the rules which govern this process.

The most straightfoward pattern of ablaut is found in monosyllabic roots of the form /*CaC-/ in which low central vowel /a/ is fronted to low mid front /ɛ/ or low mid back /ɔ/ in harmony with the first vowel of the suffix as follows:

The root /ˈsa-/ “say/speak” follows this pattern except in the past, where /a/ is raised to /i/ due to the interpolation of subphonemic palatal glide [j] (above):

It should be noted from the outset that the stem forms designated in these sections do not in this respect necessarily apply to all suffixes in that stem class (although all forms should be predictable from Schieffelin and Feld's examples as both future imperative /-(ɺ)ɛbi/ and past /-ɔ/ appear in ablauting environments.) Unlike the realization of final segments which is governed by the initial consonant of the suffix, which is the organizing principle behind stem form assignations, ablaut is governed by the first vowel of the suffix in combination with the vowel of the monosyllabic stem, or the penultimate vowel of disyllabic stems (below.) For example, the root vowel in stem III forms for /ham-/ “go” varies as follows (q.v. Grosh and Grosh 2004a: 10, 2004b: 71):

(n.b. Schieffelin and Feld give the second and third person present form of “go” as /hɛn-ab/; we follow Grosh and Grosh as their attestation makes more phonologial sense.)

… :

…:

This phenomenon also affects disyllabic verb roots of the form /ˈ(C)aCe ˈ(C)aCa/ when the second vowel is regularly lost in stem III and stem IV forms to yield a monosyllabic stem /(C)aC/:

One example which unexpectedly takes /-ɛ/ in the past varies accordingly:

…:

…:

…:

Disyllabic roots of the form /ˈ(C)iCa- ˈ(C)uCa- ˈ(C)aCi- ˈ(C)aCu-/ undergo vowel alternations in the stem:

…:

Roots of the form /(C)aˈCi- (C)aˈCu-/ do not undergo vowel alternation and maintain their final vowel in both stem III and stem IV forms, showing the distinction to have originated as a difference in stress. There are no examples /(C)iˈCa- (C)uˈCa-/ because /a/ cannot take final stress (above):

…:

…:

…:

…:

A small number of otherwise monosyllabic roots have an unexplained stressed rounded high back vowel /u/ appended to the stem I and stem IV forms while stem II and III forms remain monosyllabic and undergo fronting ablaut (above); cf. /as/ “netbag;” these are distinct from disyllables /ˈ(C)aCu-(C)aˈCu-/ (above):

It has been mentioned earlier that, following a monosyllabic root with initial or final bilabial nasal /m/, imperative /-ma/ is expressed as /-na/:

Several disyllables with initial /m/ follow a different pattern, with /-ma/ being expressed as merely /-a/, most likely following a stage in which it was /-na/:

…:

In addition to the paradigms presented above, there are a small number of unambiguously suppletive paradigms associated with very basic meanings. Following is an exhaustive list of those given in Schieffelin and Feld (1998):

Most of what would be considered adjectives in English are formally verbs in Kaluli. Their paradigms are most often underspecified in Schieffelin and Feld; nevertheless they can in many instances be definitively assigned to one of the classes presented above. What distinguishes them in practice is not adjectival meanings per se, but the lack of an attested imperative (i.e. stem I) form, which is the usual citation form of verbs in Schieffelin and Feld's dictionary. Stem II forms are typically exemplified with second/third person present /-(ɺ)ab/ rather than future imperative /-(ɺ)ɛbi/ and with second/third person future /-mɛib/ rather than first person future /-mɛnɔ/. Stem I forms are never exemplified so are not shown here:

(A small number of examples of first person present /-(ɺ)ɔɺ/ and past progressive /-(ɺ)ale/ are not included here because they are in all instances redundant to those with /-(ɺ)ab/ second/third person present. Where a stem II form with second/third person future /-mɛib/ is shown in brackets, this is because first person future /-mɛnɔ/ was given instead.)

One example is indeterminate between stressed root-final /ˈdV-/ and unstressed /dV-/ because retention of apical voiced stop /d/ in stem II vs. lenition to non-stop /ɺ./ is disagnostic for this distinction; alternation between them appears to be irrregular:

Where no step II form is attested, it's not possible to determine whether a final vowel was present following apicals /n d s ɺ/:

Wher no stem II form is attested, it is also not possible to distinguish unstressed stem-final /dV-/ from stressed /ˈdV-/, as both regularly retain [d] in stem II and stem III:

… auxilliaries and compounds ….

… auxilliary /-fo-/ … Schieffelin and Feld (1998: xii) “often indicates complete, thorough or full action.” The inflectional pattern of /-fo-/ is unique and instantly recognizable, differing from that of disyllabic verb roots with final /fV/ (above) in both stem II and stem III forms.

…past tense is signified by /-ɛ/, as it is on monosyllabic verb roots with final bilabials /b f/, rather than /-ɔ/, suggesting that /-fo-/ has been reduced to /-f-/ [f ø] as it has been in stems II and III:

… :

One of these examples shows an apparently irregular reduction in the form of the root in the past:

… /-ɛ-fo₁-/ … literal meaning “put down and leave there” or the more figurative “completely”. Following low central and low mid back /a ɔ/, low mid front vowel /ɛ/ is raised to high front /i/ [-i-fo-], with /i/ retained in the past where /ɛ/ is dropped … vowel sequences /aɛ ɔɛ/ do not occur …:

…:

Ccmpounds with the second member /aɺi-fo-/ also follow this pattern, even though independent occurences of /aɺi-fo-/ take past tense final /-ɔ/ (below):

In one example, final high mid front vowel /e/ of the root is unexpectedly given as high front [ij] when followed by /aɺi-fo₁-/:

If the final vowel of the stem I form is low mid front /ɛ/ or low central /a/, the vowel is dropped when followed by /-aɺi-fo₁-/:

In the more common pattern, the past tense is indicated by the suffix /-ɔ/, suggesting that the inflections of auxilliary /-fo-/ are in the process of being regularized:

One example is indeterminate between /-fo₁-/ and /-fo₂-/ conjugations because no stem IV (past) form is attested:

…/-ɛ-fo₂-/ … Besides the difference in the vowel of the past which distinguishes /-fo₂-/ from /-fo₁-/, this pattern is analogous to that of /-ɛ-fo₁-/ (above,) in which preceding low and low mid back vowels /a ɔ/ induce the raising of underlying /ɛ/ to high front [i]. We can then urther generalize the rule to include both preceding low mid vowels /æ ɔ/ as well as low central /a/; thus this construction would be underlyingly /-ɛ-fo₂-/ with the phonetic realization [-i-fo-]:

…/-gɛ-ɛ-fo₂-/ … [-gɛ-i-fo-] …:

The auxilliary /-ˈfeɺe-/ seems to function as a pluralizing counterpart of /-fo-/ (above,) and is likewise not associated with any independent root. Its stem forms are those of a regular disyllabic root:

…:

In one of these examples, the final vowel of stem I and stem II forms appears to have been influenced by that of the auxlliary /-ɺɛ-/ (below):

… /-gɛ-ɛ-ˈfeɺe-/ realized as [-gɛ-i-feɺe-] due to the regular raising of /ɛ/ when preceded by low mid and low vowels /ɛ ɔ a/, is analogous to the complex auxilliary /-gɛ-ɛ- fo₂ -/ [-gɛ-i-fo-] (above):

One example given as merely /-gɛ-ˈfeɺe-/ is likely in error for [gɛ-i-ˈfeɺe-], since there are no other examples of either /-gɛ-ˈfeɺe-/ or /-gɛ-fo₂-/:

… /-aɺi-ˈfeɺe-/ …:

As with /aɺi-fo₁-/ (above,) if the final vowel of the stem I form is low mid front /ɛ/ or low central /a/, the vowel is dropped when followed by /-aɺi-ˈfeɺe-/:

The pattern of forms suffixed with auxilliary /-ɺɛ-/ differs from that of stressed root-final /ˈɺV-/ (above) in collapsing future imperative underlying /ɺɛ-ɺɛbi/ to [ɺ-ɛbi], and from unstressed root-final /ɺV-/ (above) in the past tense, in which final /ɛ/ is retained while past /-ɔ/ is dropped; thus it might be viewed as midway between stressed and unstressed:

The most typical function of /ɺɛ-/ is to pluralize the object rather than the subject of a verb. Other examples have no obvious pluralizing function, suggesting perhaps a continuative or iterative sense:

In four examples, auxilliary vowel /ɛ/ is unexpectedly given as high-mid /e/:

…:

…:

…:

… auxilliary /-tʰanˈden-/ [-tʰanden- -sanden- -sanen-] … Schieffelin and Feld (1998: xii.) … “often indicates many people doing repeated action to many things.” These observe the same pattern of conjugation as do other verbs with final apical nasal /n/ (above):

…:

A great many Kaluli verb stems result from compounding, in which verb roots are free to compile in what are in essence zero-marked serial constructions. In principle any two or more verbs can be joined to form compounds. Nominals may also serve as first members of verbal compounds. First constituents generally take their stem I forms, with consonant-final roots interpolating a copy vowel between them and the second constituent. Besides semantics, compounds may often be identified when their suffixation patterns are those of monosyllabic or suppletive paradigms. Some examples of these compounds follow. In a few instances the meaning of the first constituent is not known.

The most common second member of a verbal compound is /ˈdi-/ “take,” to the point where it might well be considered an auxilliary. Realized with an apical non-stop /-ɺi-/, it is distinguishable from multisyllabic roots with final syllables /di ɺi/ because the resulting compounds follow the the inflectional pattern of monosyllable /ˈdi-/:

…:

…

Another second member of compounds which might be viewed as an auxilliary is /ˈdo-/ “be/stay/do.” It is found most conspicuously on terms which indicate colors. As with auxilliary /-ˈdi-/ [-ɺi-] “take” above, it is generally realized with an apical non-stop /-ɺo-/, but this is not quite consistent with several examples showing an alternation between [ɺo do]. It is distinguishable from multisyllabic roots with final syllables /do ɺo/ because the resulting compounds follow the the inflectional pattern of monosyllable /ˈdo-/:

…:

One apparently irregular example is attested only in a compound with /-ˈfeɺe-/ (above;) here the base form is /son-do-/ rather than the expected /sona-ɺo-/:

Compounds with second member /ˈmi-/ “give”, like those involving /ˈdi-/ “take” above, are conjugated just as is /ˈmi-/ when occuring independently:

…:

…. irregular verbs /ˈme-/ “come”, /ˈmese-/ “sit”, /ˈmeɺe-/ “stand” and /ˈham-/ “go” …

… /ˈme-/ “come” …:

…:

…:

One example of this pattern is attested only in the third person present:

…. /ˈmese-/ “sit” … semantics ranging from actual sitting to a continuative sense:

…. “sit” … auxilliary /-fo-/ (above) …:

A second less common pattern uses an alternative form of “sit” in the past tense, /asi/, which as found also as the base of /asi-fo₂-/ “sit awhile” (above.) As /asi/ is likewise given as an alternative past form alongside /asi-f-ɔ/, we can discern that this must be analyzed as past /asi-ø/ of an independent root /aˈsi-/ with final stress, as a stem IV form would not normally serve as the base of a compound with auxilliary /-fo-/:

…. /ˈmeɺe-/ “stand” …

… This /ɛ/ is most likely identical to Grosh and Grosh's (2004a: 52, 59-60) proclitic /ɛ-/ indicating repeated action:

…. “stand” … auxilliary /-fo-/ (above) …

…. /ham-/ “go” …

The semantic import of the addition of first member /ˈdi-/ “take” is quite specialized. Its primary function is to add a transitive sense to otherwise intransitive verbs, or a ditransitive sense to those which are already inherently transitive. The verb /di-kʰiˈdɛn-/ for example is defined by Schieffelin and Feld as “tell one person to wake up another.” For whatever reason, most of those with interpersonal meanings, notably excepting /di-ˈmi-/ “give,” involve shame or derision:

… /di-ɛ-/ …:

The paradigm for /di-ɛ-ˈham-/ “take away” appears to be irregular in the past such that another particle is needed to explain it. Schieffelin and Feld (1998: 31, 32) gloss /di-ɛ-gɛ-ø/ as “having taken;” we assume this to be the base of the form /di-ɛ-g-ane-ø/. Contra Schieffelin and Feld, the element /-gɛ-/ cannot be the present in the base of other compounds with prefixed /di-ɛ-/:

An element /de-/ of unknown origin and meaning is found prefixed to a small number of roots involving motion or changes of position:

Other compounds include the following:

In a small number of examples which might be assumed to be compounds due to the appearance of voiceless stops [tʰ kʰ] in medial position, the meaning of the individual constituents is not known:

… tone … minimal pairs …

Differences in tone patterns of the roots can carry over to the auxilliary:

…:

…:

Schieffelin and Feld (1998: xvii) … nominals which are derived from the past forms of verbs, which they state are found only utterance-finally:

…