Parts of speech in Yawarana

TBD: Introduction

Distinguishing parts of speech

  • Koehn and Koehn 1986: 111 on Apalaí: "Particles follow words of any class other than the ideophone, and never occur as free forms or in isolation."

Verbs

TBD

Nouns

TBD

Adverbs

  • copredicative function
  • no person inflection
  • deriving aderbs: -kepropto do: is -ke negated with -jra only when on noun roots?

Postpositions

TBD

Shared morphology

  • third person prefixes & linker are (partially) shared between nouns and postpositions
  • -jpëpst’ on nouns, pronouns & verbs

Pronominal clitics

to do: expand this section Transitive and verbs, nouns, and postpositions share preposed sap person markers. The occurrence of bound first person u= on members of all four parts of speech is illustrated in ; illustrates the same distribution for second person =. The 1+2 form ej= is very rare compared to the first and second person forms.

    1. Who was going to feed me then? / ¿quién más me iba a alimentar pues?
    1. You're going to go like this? Don't go like this. / así tu te vas a ir, no puedes ir así
    2. I will beat you to death. / te voy a matar a golpes
    3. Didn't he say it directly to you? / ¿no te lo dijo directamente a tí?

When occurring on nouns and adpositions, these elements behave identically to a free pronoun by triggering the linker y-. An alternative analysis in which the person markers are prefixes would require the y occurring on nouns and postpositions to be part of the prefix, so e.g. më- / _C, mëy- / _V. However, no such y occurs on verbs . Such an analysis would therefore need to postulate that there are phonologically conditioned allomorphs on nouns and postpositions, but not on verbs. An analysis as cliticized pronominal forms does not face that issue, since the absence of y on verbs is expected by the absence of a linker on verbs.

Other applicable criteria point to a clitic analysis, too (ref).

Comparison of phonologically bound person markers.

Person Form Parts of speech Free counterpart y-? Stem-based allomorphy?
1 u- Vt/Vi/N/Postp wïrë + -
2 - Vt/Vi/N/Postp mërë + -
2 a- N - + +
3 ta- Vt tëwï? - -
3 i- N/Postp - - +
3 t- N/Postp - - +
3 - Postp tëwï - +

Interestingly, second person is much more frequently expressed by a bound form (ref).

Ratio of bound person markers compared to free pronouns.

1 2 1+2
Vt 5.04% (7/139) 93.10% (27/29) 0.00% (0/35)
Vi 5.60% (7/125) 97.06% (33/34) 3.33% (2/60)
N 14.61% (58/397) 95.40% (83/87) 2.35% (5/213)
Postp 9.38% (3/32) 100.00% (90/90) 0.00% (0/22)

Ratio of bound person markers compared to free pronouns, split by initial segment.

1 2 1+2
Vt / _C 7.53% (7/93) 100.00% (27/27) 0.00% (0/29)
Vt / _V 0.00% (0/46) 0.00% (0/2) 0.00% (0/6)
Vi / _C 0.00% (0/109) 97.06% (33/34) 0.00% (0/54)
Vi / _V 43.75% (7/16) 33.33% (2/6)
N / _C 2.91% (9/309) 96.51% (83/86) 0.00% (0/170)
N / _V 55.68% (49/88) 0.00% (0/1) 11.63% (5/43)
Postp / _C 0.00% (0/25) 100.00% (90/90) 0.00% (0/21)
Postp / _V 42.86% (3/7) 0.00% (0/1)

Derivation and productivity

  • productive class-changing process w/ lexically conditioned suffixes
  • semantic variation & non-compositional meanings
  • some constructions simply need a different word class, no meaning change per se