Teleology and generics

This repository contains the experiments, data, analyses, and figures for the paper "Teleology and generics," by David Rose, Siying Zhang, Qi Han and Tobias Gerstenberg.

The preprint can be found here

Contents:

• Introduction
• Repository structure

Introduction

Generic statements, such as ``Bees are striped'' are thought to be a central vehicle by which essentialist beliefs are transmitted. But work on generics and essentialism almost never focuses on the type of properties mentioned in generic statements. We test the hypothesis that teleological properties, what something is for, affect categorization judgments more strongly than behavioral, biological, or social properties. In Experiment 1, participants categorized properties as being either behavioral, biological, social, or teleological. In Experiment 2, we used the top four properties from each group to describe a generic noun or a specific individual. Participants then categorized creatures that had one of their properties transformed. We found that changes to teleological properties had the strongest impact on categorization judgments. In Experiment 3, we also found that teleological properties mattered more in an induction task. We suggest teleological properties play this privileged role in categorization because they are treated as essential properties.

Repository structure

├── code
│   ├── R
│   ├── experiments
│   │   ├── experiment1
│   │   ├── ...
│   └── python
├── data
│   ├── experiment1
│   ├── experiment2
│   └── experiment3
├── docs
│   ├── experiment2
│   └── experiment3
├── figures
│   ├── experiment1
│   ├── experiment2
│   ├── experiment3
└── writeup

• code/ contains all the code for the experiments, analyzing data and generating figures.
  • experiments contains code for each experiment that was run. Pre-registrations for all experiments may be accessed via the Open Science Framework here
    • experiment1 (pre-registration) was run in psiturk
    • experiment2 (pre-registration) and experiment3 (pre-registration) were run in jsPsych. These are copied in docs below where they can be rendered in your browser.
  • python contains scripts that were used to get the probabilities of sentence completions for properties used in experiment 2.
  • R contains the analysis scripts that were used to analyze data and generate figures (view a rendered file here).
• data/ contains anonymized data from all experiments:
  • experiment1 contains experiment1.db which includes trial and demographic data.
  • experiment2 contains experiment2.csv which includes the trial data and experiment2_demographics.csv which includes demographic data. The remaining files are for getting the probabilities of sentence completions from large language models for the properties used in experiment 2.
  • experiment3 contains experiment3.csv which includes the trial data for experiment 3 while experiment2_demographics.csv contains the demographic data.
• docs/ contains all the experiment code for experiment2 and experiment3. You can preview the experiments below:
  • Experiment 2 - Generic condition, Specific condition
  • Experiment 3 - Generic condition, Specific condition
• figures/ contains all the figures from the paper (generated using the script in code/R/).