Life is a Circus and We are the Clowns - Automatically Finding Analogies between Situations and Processes

Sep 14, 2026

Analogy-making enables reasoning, abstraction, flexible categorization, and counterfactual inference, capabilities still missing in today’s leading AI systems. Prior NLP work mainly targets simple word analogies and often depends on manually engineered representations, limiting scalability to richer relational cases. This study automatically extracts entities and relations from pairs of natural-language procedural texts and aligns domains via relational similarity. The proposed interpretable method achieves 87% mapping accuracy on procedural texts and 94% on cognitive-psychology stories, with 79% precision and robustness to paraphrasing.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Abstract

1 Introduction

Motivation and background

Problem setting and goals

Related work (Structure Mapping Theory)

Proposed approach and examples

Life is a Circus and We are the Clowns: Automatically Finding Analogies between Situations and Processes Oren Sultan The Hebrew University of Jerusalem [oren.sultan@mail.huji.ac.il](oren.sultan@mail.huji.ac.il) Dafna Shahaf The Hebrew University of Jerusalem [dshahaf@cs.huji.ac.il](dshahaf@cs.huji.ac.il) Abstract Analogy-making gives rise to reasoning, abstraction, flexible categorization and counterfactual inference – abilities lacking in even the best AI systems today. Much research has suggested that analogies are key to non-brittle systems that can adapt to new domains. Despite their importance, analogies received little attention in the NLP community, with most research focusing on simple word analogies. Work that tackled more complex analogies relied heavily on manually constructed, hard-to-scale input representations. In this work, we explore amore realistic, challenging setup: our input is a pair of natural language procedural texts, describing a situation or a process (e.g., how the heart works/how a pump works) . Our goal is to automatically extract entities and their relations from the text and find a mapping between the different domains based on relational similarity (e.g., blood is mapped to water) . We develop an interpretable, scalable algorithm and demonstrate that it identifies the correct mappings 87% of the time for procedural textsand 94% for stories from cognitive-psychology literature. We show it can extract analogies from a large dataset of procedural texts, achieving 79% precision (analogy prevalence in data: 3%) . Lastly, we demonstrate that our algorithm is robust to paraphrasing the input texts 1. 1 Introduction The ability to find parallels across diverse domainsand transfer ideas across them is one of the pinnacles of human cognition. The analogous reasoning process allows us to abstract information, form flexible concepts and solve problems based on our previous experience (Minsky, 1988 ; Hofstadter and Sander, 2013 ; Holyoak, 1984)– abilities that current AI systems still lack. Many researchers have 1Data and code are available in our GitHub repository [https://github.com/orensul/analogies_mining](https://github.com/orensul/analogies_mining) suggested that analogy-making is essential for engineering non-brittle AI that can robustly generalize and adapt to diverse domains (Mitchell, 2021b) . Surprisingly, despite analogy’s important role in the way humans understand language, the problem of recognizing analogies has received relatively little attention in NLP. Most works have focused on SAT-type of analogies (“a to b is like c to d”), with recent works (Linzen, 2016 ; Ushio et al., 2021 ; Brown et al., 2020) showing that models still struggle with abstract, loosely defined relations. In this work, we focus on a different type of analogies: analogies between situations or processes. Here the input is two domains (e.g., heart and pump) and the goal is to map objects from the base domain to objects from the target domain. Importantly, the mapping should rely on a common relational structure rather than object attributes, making it challenging for NLP methods. The most influential work in this line of re search is Structure Mapping Theory (SMT) (Gentner, 1983) . In SMT and much of its follow-up work (Falkenhainer et al., 1989 ; Turney, 2008 ; Forbus et al., 2011), domain descriptions are sets of statements in a highly structured language, e.g., CAUSE(PULL(piston),CAUSE(GREATER(PRESSURE(water), PRESSURE(pipe)),FLOW(water,pipe))) Many have argued that too much human creativity is required to construct these inputs, and the analogy is already expressed in them (Chalmers et al., 1992) . They are also brittle and hard-to-scale. In our work, we explore a more realistic, challenging setup. Our input is a pair of two procedural texts, describing a situation or a process, expressed in natural language. We develop an algorithm to automatically extract entities and their

FAQs

What problem does the paper address in analogy mining?

It studies recognizing analogies between situations or processes described in natural language procedural texts, focusing on mappings based on relational structure rather than object attributes.

How does the proposed method find mappings between two domains?

It automatically extracts entities and relations from each text, then finds a mapping by comparing relational similarity and the roles entities play in the descriptions.

What performance results does the paper report?

The method identifies correct mappings 87% of the time for procedural texts and 94% for stories from cognitive-psychology literature, while reaching 79% precision on a large dataset and remaining robust to paraphrasing.

Sep 14, 2026

Uploaded by Taylor Morgan