Add speciesist language linter (woke integration or dedicated)

[stuckvgn]

Welcome

• I searched for existing GitHub issues

Is your feature request related to a problem? Please describe.

golangci-lint currently has no linter for inclusive language. Other ecosystems have tools like woke (Go-based, MIT, 500+ stars) and alex that detect non-inclusive terminology in source code — covering racial, gender, and ableist language.

However, none of the existing inclusive language tools cover speciesist language: phrases that normalize violence toward animals, particularly in technical contexts. This is relevant to Go codebases because:

• Comments and documentation frequently use idioms like "kill two birds with one stone", "beat a dead horse", "more than one way to skin a cat"
• Technical terminology in Go code uses animal exploitation metaphors: "monkey patching" (runtime patching), "duck typing" (structural typing/interfaces), "canary deployment" (progressive rollout), "cattle vs. pets" (ephemeral vs. persistent infrastructure)
• Identifier naming may use "master/slave" patterns that several Go projects have already moved away from

Research backing this: Takeshita, Rzepka & Araki (2022) documented speciesist language patterns in code in Information Processing & Management. Hagendorff, Bossert, Tse & Singer (2023) documented speciesist bias in AI systems in AI and Ethics (DOI: 10.1007/s43681-023-00380-w).

The Inclusive Naming Initiative (founded by CNCF/Kubernetes maintainers) has already driven terminology changes in the Go ecosystem — this extends that work to a category not yet covered.

Describe the solution you'd like.

A go/analysis-based linter that detects speciesist language in Go source code (comments, string literals, identifiers, documentation). Two possible approaches:

Option A: Wrap woke's detection logic
get-woke/woke is already written in Go and MIT-licensed. Its rule matching engine could be adapted into a go/analysis analyzer. The Open-Paws/speciesist-language-scanner project provides woke-compatible rule files specifically for speciesist terminology with curated alternatives.

Option B: Dedicated go/analysis linter
A new analyzer (similar in approach to misspell or dupword) that scans comments, string literals, and identifiers for speciesist terms and suggests alternatives. This could be built as a standalone go/analysis module that golangci-lint integrates.

What it would flag (Go-specific examples):

// TODO: kill two birds with one stone by refactoring both handlers
// -> "accomplish two things at once"

// This is a wild goose chase, the bug is elsewhere  
// -> "futile search"

// monkey-patch the HTTP client for testing
// -> "runtime patch"

type MasterSlaveReplication struct { ... }
// -> "PrimaryReplicaReplication"

Configuration should allow:

• Severity levels (error/warning/info) per category
• Custom term additions and overrides
• Category filtering (e.g., enable only tech terminology, disable idioms)
• Scope control (comments only, identifiers only, string literals, etc.)

Describe alternatives you've considered.

1. Running woke separately — This works but misses golangci-lint's Go-aware AST context (distinguishing comments from identifiers from string literals), unified configuration, IDE integration, and nolint directive support.

2. Module plugin — The golangci-lint module plugin system allows adding custom linters without upstream changes. A speciesist language analyzer could be distributed as a module plugin. However, upstream inclusion would give this much broader adoption and signal that the Go ecosystem takes this category of inclusive language seriously.

3. Vale/pre-commit hooks — Tools like Vale can scan documentation but don't understand Go syntax, can't distinguish between code and comments, and aren't integrated into the Go development workflow.

Additional context.

Existing rule sets are ready to use:
The Open-Paws/speciesist-language-scanner repository provides curated rule files for woke, Vale, and alex covering 25+ speciesist terms across three categories:

• Violent animal idioms (kill two birds, beat a dead horse, skin a cat, etc.)
• Animal-as-object metaphors (guinea pig, sacred cow, scapegoat, etc.)
• Technical terminology (cattle vs. pets, canary deployment, monkey patch, duck typing, etc.)

Each rule includes severity levels, word boundary settings, and academically sourced notes explaining why the term is problematic and why the alternative is more precise.

This fits golangci-lint's existing pattern:

• misspell already does text-level analysis across comments and identifiers
• dupword analyzes comment text for repeated words
• godot enforces comment style
• gosmopolitan flags internationalization issues

A speciesist language linter would be a natural addition to this text-analysis category.

Open Paws (openpaws.ai) is a 501(c)(3) nonprofit building AI infrastructure for animal advocacy organizations. We're happy to contribute the implementation or collaborate on it.

[boring-cyborg[bot]]

Hey, thank you for opening your first Issue ! 🙂 If you would like to contribute we have a guide for contributors.