@inproceedings{sun-kazakov-2026-kinship,
title = "Kinship Data Benchmark for Multi-hop Reasoning",
author = "Sun, Tianda and
Kazakov, Dimitar",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.274/",
pages = "5767--5784",
ISBN = "979-8-89176-462-0",
abstract = "Multi-hop kinship reasoning is a natural testbed for LLM compositionality, but existing benchmarks (notably CLUTRR) cover only the descriptive Eskimo system. We introduce KinshipQA, a procedurally-generated benchmark covering seven anthropologically-documented kinship systems (Eskimo, Sudanese, Hawaiian, Iroquois, Dravidian, Crow, Omaha) and up to six reasoning hops, with a tunable simulator horizon that eliminates exact-instance pretraining overlap. Evaluating six LLMs, we find a 40.9{\%} accuracy drop when reasoning shifts from biological multi-hop to culturally-marked classification on the five non-descriptive systems. The drop follows the anthropological hierarchy descriptive {\ensuremath{>}} bifurcate {\ensuremath{>}} generational {\ensuremath{>}} skewing, persists under chain-of-thought and few-shot prompting, and compounds with depth (10.6{\%} at 5{--}6 hops). Humans given the override rule reach 89.0{\%} versus 50.7{\%} for LLMs, placing the failure in the model rather than the task. Two follow-up experiments suggest distinct contributors. A fictional-rule control swapping system labels and kin terms for invented strings raises accuracy by 6.1{\%}, implicating familiar English surface forms. An in-context-rule probe prepending the override rule helps skewing systems (+17.1{\%}) but hurts non-skewing systems whose baseline already exceeds {\textasciitilde}60{\%} (-13.4{\%}), consistent with a missing skewing prior alongside rule interference where the model already has a working approximation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sun-kazakov-2026-kinship">
<titleInfo>
<title>Kinship Data Benchmark for Multi-hop Reasoning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Tianda</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dimitar</namePart>
<namePart type="family">Kazakov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Multi-hop kinship reasoning is a natural testbed for LLM compositionality, but existing benchmarks (notably CLUTRR) cover only the descriptive Eskimo system. We introduce KinshipQA, a procedurally-generated benchmark covering seven anthropologically-documented kinship systems (Eskimo, Sudanese, Hawaiian, Iroquois, Dravidian, Crow, Omaha) and up to six reasoning hops, with a tunable simulator horizon that eliminates exact-instance pretraining overlap. Evaluating six LLMs, we find a 40.9% accuracy drop when reasoning shifts from biological multi-hop to culturally-marked classification on the five non-descriptive systems. The drop follows the anthropological hierarchy descriptive \ensuremath> bifurcate \ensuremath> generational \ensuremath> skewing, persists under chain-of-thought and few-shot prompting, and compounds with depth (10.6% at 5–6 hops). Humans given the override rule reach 89.0% versus 50.7% for LLMs, placing the failure in the model rather than the task. Two follow-up experiments suggest distinct contributors. A fictional-rule control swapping system labels and kin terms for invented strings raises accuracy by 6.1%, implicating familiar English surface forms. An in-context-rule probe prepending the override rule helps skewing systems (+17.1%) but hurts non-skewing systems whose baseline already exceeds ~60% (-13.4%), consistent with a missing skewing prior alongside rule interference where the model already has a working approximation.</abstract>
<identifier type="citekey">sun-kazakov-2026-kinship</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.274/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>5767</start>
<end>5784</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Kinship Data Benchmark for Multi-hop Reasoning
%A Sun, Tianda
%A Kazakov, Dimitar
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F sun-kazakov-2026-kinship
%X Multi-hop kinship reasoning is a natural testbed for LLM compositionality, but existing benchmarks (notably CLUTRR) cover only the descriptive Eskimo system. We introduce KinshipQA, a procedurally-generated benchmark covering seven anthropologically-documented kinship systems (Eskimo, Sudanese, Hawaiian, Iroquois, Dravidian, Crow, Omaha) and up to six reasoning hops, with a tunable simulator horizon that eliminates exact-instance pretraining overlap. Evaluating six LLMs, we find a 40.9% accuracy drop when reasoning shifts from biological multi-hop to culturally-marked classification on the five non-descriptive systems. The drop follows the anthropological hierarchy descriptive \ensuremath> bifurcate \ensuremath> generational \ensuremath> skewing, persists under chain-of-thought and few-shot prompting, and compounds with depth (10.6% at 5–6 hops). Humans given the override rule reach 89.0% versus 50.7% for LLMs, placing the failure in the model rather than the task. Two follow-up experiments suggest distinct contributors. A fictional-rule control swapping system labels and kin terms for invented strings raises accuracy by 6.1%, implicating familiar English surface forms. An in-context-rule probe prepending the override rule helps skewing systems (+17.1%) but hurts non-skewing systems whose baseline already exceeds ~60% (-13.4%), consistent with a missing skewing prior alongside rule interference where the model already has a working approximation.
%U https://aclanthology.org/2026.findings-emnlp.274/
%P 5767-5784
Markdown (Informal)
[Kinship Data Benchmark for Multi-hop Reasoning](https://aclanthology.org/2026.findings-emnlp.274/) (Sun & Kazakov, Findings 2026)
ACL
- Tianda Sun and Dimitar Kazakov. 2026. Kinship Data Benchmark for Multi-hop Reasoning. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 5767–5784, Budapest, Hungary. Association for Computational Linguistics.