2025-04-02 chatgpt claude # chatgpt version Calling **non-procedural knowledge** _declarative knowledge_ is **partially accurate**—but also **incomplete** and **potentially misleading** depending on the context and level of precision required. Here’s a breakdown: --- ### ✅ **Where the Terms Overlap** **Declarative knowledge** typically refers to **“knowing that”**—facts, definitions, concepts, and truths that can be stated explicitly. This includes: - Factual knowledge: “Paris is the capital of France” - Conceptual knowledge: “Photosynthesis is how plants make energy” - Semantic knowledge: “A mammal is a warm-blooded vertebrate” These are core subsets of **non-procedural knowledge**, which makes the label _declarative_ useful as a broad-strokes term. --- ### ❌ **Where Declarative Knowledge Falls Short** **Non-procedural knowledge** is a **much broader category** than declarative knowledge. It includes: - **Causal knowledge** ("Pulling this lever triggers a shutdown") - **Contextual knowledge** ("The word 'charge' means something different in physics than in law") - **Epistemological knowledge** ("I’m only 60% confident in this information") - **Normative knowledge** ("It’s unethical to fire an employee without warning") - **Strategic knowledge** ("To maximize revenue, prioritize high-LTV customers") These types: - May not be **explicitly declarable** in a single sentence - Often involve **conditionality**, **judgment**, or **meta-awareness** - Cannot be cleanly formalized into simple “knowing that” statements In other words: _non-procedural knowledge includes declarative knowledge, but also spans well into cognitive, ethical, causal, and strategic territories that declarative knowledge doesn’t fully capture_. --- ### 🧠 Better Framing If you're looking for precision, consider the following hierarchy: - **Declarative Knowledge** ⊂ **Non-Procedural Knowledge** Non-procedural = declarative + causal + contextual + strategic + normative + epistemological + experiential + meta-cognitive... --- ### 🧩 When Is It Okay to Conflate Them? |**Context**|**Calling Non-Procedural = Declarative**|**Why**| |---|---|---| |High-level intro / informal discussion|✅ Acceptable|Helps distinguish from procedural knowledge clearly| |Technical/philosophical conversation|❌ Inaccurate|Ignores critical dimensions like causality, ethics, strategy| |Ontology design / knowledge engineering|❌ Misleading|Requires fine-grained distinctions for modeling| |Teaching AI concepts to non-technical audiences|✅ With caveats|“Non-procedural = knowledge that can be stated, not acted out” works if you add nuance later| --- ### ⚖️ Final Verdict: Calling non-procedural knowledge _declarative knowledge_ is like calling all square shapes _rectangles_. Technically correct in one direction—but conceptually limiting in the other. **Use "declarative knowledge" as a subset, not a synonym.** If you're building or describing advanced AI systems, the richer term—**non-procedural knowledge**—captures far more of the cognitive landscape. # claude version Calling non-procedural knowledge "declarative knowledge" is generally accurate, but with some important nuances to consider. The terms are often used interchangeably in cognitive science and AI research, as they both refer to knowledge about facts, concepts, principles, and relationships that can be explicitly stated or "declared" - in contrast to procedural knowledge which involves knowing how to perform actions or tasks. However, there are some considerations that make this equivalence not perfectly precise: 1. **Scope differences**: Declarative knowledge is typically considered a subset of non-procedural knowledge in many taxonomies. While declarative knowledge refers specifically to facts and concepts that can be explicitly stated, non-procedural knowledge might also encompass more implicit forms of knowledge like experiential, contextual, or phenomenological knowledge that aren't easily declared. 2. **Granularity issues**: Non-procedural knowledge, as discussed in your documents, includes multiple categories with varying characteristics (factual, conceptual, relational, epistemological, normative, etc.). Declarative knowledge traditionally focuses more on the factual and conceptual categories. 3. **Philosophical distinctions**: Some philosophers and cognitive scientists make distinctions between "knowing that" (declarative) and other forms of non-procedural knowledge like "knowing of" (familiarity) or forms of tacit knowledge that cannot be fully articulated. 4. **Implementation differences**: In AI systems, declarative knowledge is often implemented in specific ways (like semantic networks or logical assertions), while non-procedural knowledge might be represented through various methods including distributed representations that don't map cleanly to explicit declarations. For general purposes and most practical discussions, using "declarative knowledge" as a synonym for "non-procedural knowledge" works well and would be understood by most experts in the field. The equivalence is accurate enough for most contexts, especially when contrasting with procedural knowledge. If you're writing for a specialized audience or need precise terminology, you might want to clarify that you're using declarative knowledge in its broader sense to encompass the full spectrum of non-procedural knowledge types.