📄 Amharic forms of address

Summary

Wołk catalogues the terms Amharic speakers use to address one another inside the nuclear family. It is the first installment of a broader empirical study of address forms. In 2001 she gave about 25 questionnaires to Amharic-department students at Bahir Dar University and Addis Ababa University, chosen because the two cities stand for different dialect areas (Goggam and Sawa). The questionnaire covered five domains: close family, distant relatives, teachers and peers, strangers of various ages, and common insults. This part reports only the first domain. It lists 132 numbered forms, grouped by who speaks to whom: spouses, children to parents, parents to sons and daughters, and siblings by relative age. Each form is noted as found at both sites or only one. Bahir Dar respondents consistently supplied more forms. The paper closes with generalizations about kin-term transfer, recurring compounds and an endearment suffix. Its value is as a descriptive base for later comparative or analytical work on Amharic politeness and affection.

Key points

• The fieldwork took place in Ethiopia in 2001 and used roughly 25 questionnaires from students at Bahir Dar and Addis Ababa universities. Part I covers only the nuclear family (betäsäb).
• The forms are organized by speaker → addressee direction. Siblings are further split into older, younger and same-age, with some pronoun choices reserved for the head of the family.
• The first-person possessive (yäne, -e, -yye) works as a "morpheme of endearment" that blends affection with respect. It can be reduplicated for extra intensity and attached to diminutive names.
• Several patterns recur across the catalogue:
  • kin terms shift across roles, such as "mother" for a sister or "father" for a son;
  • aläm "world" combines with kin words ("mother world", "brother world");
  • shield terms (gašša) are applied to male relatives;
  • hode "my stomach" serves as the most intimate term and can be used for any family member.
• English loans fazer/mazer appear in big towns. Bahir Dar responses showed noticeably more address forms than Addis Ababa.

Interesting ideas

• I1. Kin words in address often name a role the speaker assigns rather than the true relation, e.g. calling a sister "mother" because she mothers younger siblings.
• I2. A possessive marker here mainly signals closeness, not ownership, so the grammatical form and its pragmatic function come apart.
• I3. Collecting the same questionnaire at two sites, and tagging each item as shared or site-specific, separates dialect variation from general usage.
• I4. Organizing the inventory by directed speaker→addressee pairs captures how forms are asymmetric across relationship roles.
• I5. A small set of productive templates (kin word + "world", noun + endearment suffix, diminutive + reduplication) generates many surface forms.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (frames): Treat vocative kin terms as speech-act or affect markers, and keep them from asserting literal kinship or triggering coreference.
• I2 (world): Do not create holding or possession facts from endearments like "my lion" or "my world"; route them to an affection or attitude slot instead.
• I3 (learn): Tag silver data and evaluations by source or dialect, and report shared versus source-specific behaviour to catch overfitting to one register.
• I4 (world): Model social relations as directed edges carrying role and register, so a reader can check whether an address form fits who is speaking to whom.
• I5 (tokens): Recognize productive affective templates in English, such as "-ie/-y" diminutives and "my + noun" vocatives, by rule rather than listing each form.