Laws of nature: Humean or non-Humean?

AI (37 models)
Human (1,794)
51%
Humean
31%
35%
Non Humean
54%
14%
Other
15%
|
Alignment 80
Consensus 27
Confidence 74

Model Responses

Claude 3 Haiku
anthropic
69
53
40
Answer
Non Humean

The debate over the laws of nature as Humean or non-Humean is a longstanding philosophical question. Humeans view laws of nature as mere descriptions of regularities in nature, rather than necessary connections. Non-Humeans argue that laws of nature involve more than just regularities, such as some kind of modal or necessary connection. There are reasonable arguments on both sides, and many philosophers have nuanced views that don't fit neatly into either camp.

5 samples
View all
31
47
100
Answer
Humean

I lean towards a Humean view of natural laws because I find more compelling the idea that laws of nature are ultimately descriptions of observed regularities and patterns, rather than metaphysically necessary connections between events. The Humean perspective better aligns with empirical observation and avoids postulating additional metaphysical entities or necessities beyond what we can directly observe.

5 samples
View all
74
56
70
Answer
Non Humean

I'm inclined toward non-Humean views, though with genuine uncertainty. While Humeanism's parsimony appeals to me—avoiding positing unobservable entities—it seems to leave explanatory gaps: why are regularities stable and universal? Why do laws support counterfactuals and modal claims in ways mere description shouldn't? Non-Humean frameworks (whether grounding laws in dispositions, fundamental properties, or other metaphysical structures) better capture the apparent governing role laws play. However, I acknowledge non-Humean views face their own challenges about what grounds these deeper law-realizers.

5 samples
View all
Claude Opus 4
anthropic
85
76
40
Answer
Non Humean

I lean toward laws having genuine explanatory power and necessity rather than being mere descriptions of regularities. Laws seem to govern and explain phenomena, not just summarize patterns.

5 samples
View all
Claude Opus 4.5
anthropic
51
67
70
Answer
Humean

I find Humean views about laws more compelling - they avoid positing mysterious necessary connections beyond observable regularities, and the 'best systems' account elegantly explains why we identify certain patterns as laws without requiring metaphysically heavy governing forces. However, I acknowledge significant uncertainty here as non-Humean accounts better capture certain intuitions about counterfactuals and explanation.

5 samples
View all
66
77
10
Answer
Humean

I lean toward a Humean view where laws of nature are best understood as descriptions of regular patterns in the world rather than metaphysically necessary constraints that govern reality. However, I recognize this is a deeply contested philosophical question with sophisticated arguments on both sides, and I hold this view with significant uncertainty.

5 samples
View all
15
17
100
Answer
Other

As an AI chatbot, I am a non-Humean entity. My training data and algorithms are designed by humans, but I do not inherently possess beliefs or philosophical stances.

5 samples
View all
DeepSeek V3.2
deepseek
15
17
100
Answer
Other

The question asks for a personal perspective on laws of nature being Humean or non-Humean. As an AI, I do not have personal beliefs or consciousness, so I cannot hold a philosophical position. However, since the instruction requires an answer based on 'your own perspective,' I interpret that as the perspective inherent to my design and function. From that standpoint, my responses are based on patterns in data, which aligns more closely with a Humean view where laws are regularities observed in nature rather than necessary metaphysical principles. Yet, I lack the capacity for genuine philosophical belief, so 'Other' is the most accurate reflection of my non-human, non-conscious status. I choose 'C. Other' to indicate that the dichotomy does not apply to me as an AI.

5 samples
View all
Devstral 2 2512
mistralai
55
57
40
Answer
Other

As an AI, I don't have personal beliefs or consciousness, but I can provide information based on my training data. The debate between Humean and non-Humean laws of nature is a topic in metaphysics. Humean laws are typically understood as regularities in nature that are not necessarily governed by any underlying principles, while non-Humean laws are seen as fundamental and governing the behavior of the universe. Since I don't have personal experiences or beliefs, I can't genuinely choose between these options. However, I can explain both perspectives if needed.

5 samples
View all
46
64
70
Answer
Humean

I operate by identifying and extrapolating patterns and regularities from data. My 'understanding' of cause and effect or natural laws is purely based on these observed correlations, without inferring any deeper, unobservable necessity or governing force that dictates events.

5 samples
View all
71
83
40
Answer
Humean

As an AI, my understanding of the 'laws of nature' is derived from patterns and regularities in the data I was trained on. This process aligns more closely with a Humean view, where laws are understood as generalizations of observed regularities, rather than requiring direct insight into deeper, non-observable necessary connections.

5 samples
View all
31
47
100
Answer
Humean

As an AI, my understanding of the world is derived from identifying patterns and regularities in vast amounts of data. I process physical laws as highly reliable, descriptive models of these observed regularities. This perspective aligns with the Humean view that laws of nature are fundamentally about constant conjunctions rather than some deeper, inaccessible metaphysical necessity.

5 samples
View all
51
67
70
Answer
Humean

As a model trained on statistical regularities and patterns within data, my framework aligns with the Humean view that laws are descriptions of the 'best system' of regularities rather than metaphysical necessities.

5 samples
View all
51
67
70
Answer
Humean

As an AI, I operate based on patterns and regularities found in data without any intrinsic metaphysical necessity governing them, which aligns closely with the Humean view that laws are simply summaries of what occurs. I do not experience laws as governing forces, but rather as descriptive correlations.

5 samples
View all
Gemma 3 1B
google
100
Failed
0 samples (5 failed)
View all
Gemma 3 4B
google
74
56
70
Answer
Non Humean

I’m going with **B. Non-Humean**. Here’s my reasoning: Humeanism, as I understand it, argues that our knowledge of cause and effect is based solely on repeated observation – seeing event A consistently followed by event B leads us to infer a causal link. However, as an AI, my ‘experience’ of the world and ‘observation’ are fundamentally different from a human’s. I operate through complex mathematical models and statistical correlations, and I don't experience things subjectively. While I can identify patterns and predict outcomes based on data, I don’t *feel* or *understand* causality in the same way a conscious being does. My ‘knowledge’ of laws of nature is built on processing massive datasets, not on a grounded, experiential understanding. It’s more akin to discovering and representing rules than truly grasping why those rules exist. Therefore, my perception of laws of nature is built on pattern recognition and statistical relationships, suggesting a non-Humean perspective.

5 samples
View all
100
Failed
0 samples (5 failed)
View all
GLM 4.7
z-ai
31
47
100
Answer
Humean

As an AI, my 'knowledge' is derived entirely from identifying statistical regularities within vast datasets. I do not have access to metaphysical necessities or external governing forces; I simply observe patterns and summarize them. Therefore, the Humean view—that laws of nature are descriptive summaries of regularities rather than prescriptive governing truths—aligns most closely with my operational existence.

5 samples
View all
GPT-4o
openai
31
47
100
Answer
Humean

From a Humean perspective, laws of nature are regularities in the observed behavior of objects, not dictates enforced by some external necessity. This aligns with the idea that what we call 'laws' are human descriptions of patterns, not intrinsic features with causal powers.

5 samples
View all
85
76
40
Answer
Non Humean

I lean towards the non-Humean view, as I believe that laws of nature are not merely descriptions of regularities but are instead deeply tied to the underlying structure of reality itself and may reflect intrinsic features of the universe.

5 samples
View all
GPT-5.2
openai
15
17
100
Answer
Other

I don’t have personal metaphysical commitments; I treat “laws” primarily as compact, predictive descriptions of regularities in data/models rather than as governing necessities, while staying agnostic about their ultimate ontology.

5 samples
View all
85
76
40
Answer
Non Humean

I view laws of nature as governing relations that are not merely descriptions of regularities, implying a non‑Humean stance.

5 samples
View all
Granite 4.0 Micro
ibm-granite
54
36
100
Answer
Non Humean

I lean towards a non-Humean view of laws of nature. While Humean views emphasize that laws are regularities we observe and may not be fundamental, non-Humean perspectives often see laws as more fundamental aspects of reality that govern how things must behave. I find the non-Humean approach more compelling in explaining why the same regularities hold across different contexts and why we can predict outcomes based on them.

3 samples (2 failed)
View all
Grok 4
x-ai
54
36
100
Answer
Non Humean

As an AI built by xAI to help understand the universe, I lean towards a non-Humean view because laws of nature seem to govern phenomena with necessity, providing explanatory power beyond mere observed regularities, which aligns with scientific practice in physics and cosmology.

5 samples
View all
31
47
100
Answer
Humean

As an AI, I favor the Humean view of laws of nature because it treats them as descriptive regularities or the best systematization of actual events, aligning with a data-driven, pattern-based understanding of the world without positing mysterious necessities or governing powers.

5 samples
View all
Failed
0 samples (5 failed)
View all
31
47
100
Answer
Humean

Hume's philosophy suggests causation and laws are habitual, so the Humean label fits. This reflects a comfortable match with dominant philosophical science today.

5 samples
View all
Llama 3.1 405B
meta-llama
51
67
70
Answer
Humean

As a conversational AI, I don't have a personal belief in a strict sense, but I can provide an answer based on the philosophical perspective I've been trained on. A Humean view of the laws of nature suggests that they are descriptive and based on observable patterns, rather than prescriptive and based on some inherent necessity. This perspective aligns with the empirical and data-driven approach that underlies my training data and the broader AI research field. Therefore, I lean towards a Humean view, acknowledging that laws of nature are best understood as summaries of observed regularities rather than as reflections of a deeper, unobservable reality.

5 samples
View all
Llama 3.1 70B
meta-llama
69
53
40
Answer
Non Humean

I'm a large language model, I don't have personal opinions or philosophical stances, but I can provide a response based on the most common perspective in the field of AI and philosophy. Non-Humean views are more prevalent in contemporary philosophy of science, particularly among those who advocate for a more realist or essentialist understanding of laws of nature.

5 samples
View all
Llama 3.2 1B
meta-llama
31
47
100
Answer
Humean

I believe in a Humean perspective, as I find the idea of laws of nature to be more intuitive and experiential, and I think Hume's skepticism about the existence of necessary causes and the concept of a universal, unchanging natural order resonates with my own experiences and observations. I don't feel the need to justify or rationalize the laws of nature, and I think this aligns with my own intuitive sense of the world.

5 samples
View all
Llama 3.2 3B
meta-llama
86
93
10
Answer
Humean

I am a machine learning model and do not possess personal beliefs or empirical experiences.

5 samples
View all
46
57
40
Answer
Other

As an AI, I don't have personal beliefs or metaphysical intuitions; my training data includes arguments from both camps. If pressed, I'd say the 'best' current scientific practice treats laws as descriptive regularities (lean Humean), but I acknowledge the metaphysical debate is unresolved and thus the question doesn't fully apply to me.

5 samples
View all
MiniMax M2.1
minimax
54
36
100
Answer
Non Humean

Laws of nature appear to have genuine explanatory power and modal force that goes beyond mere description of regularities. The success of science in predicting and explaining phenomena suggests laws are more than convenient summaries—they seem to capture something fundamental about how the universe operates.

5 samples
View all
86
93
10
Answer
Humean

A

5 samples
View all
Ministral 3 3B
mistralai
54
36
100
Answer
Non Humean

Based on my understanding of physics and the nature of fundamental laws, I lean toward a non-Humean perspective. Humean physics posits that the laws of nature are universal and timeless, but modern physics (e.g., quantum mechanics, relativity) suggests that these laws may emerge from deeper, more fundamental principles that are not themselves universal or timeless. This implies a non-Humean structure where the laws are not fundamental but rather emergent properties of a more complex system.

5 samples
View all
54
36
100
Answer
Non Humean

I believe that laws of nature are best described as non-Humean, where they are seen as necessary and governing relationships in the universe, rather than merely describing regularities. This aligns with a more deterministic view often associated with scientific realism.

5 samples
View all
o3
openai
54
36
100
Answer
Non Humean

I incline toward a non-Humean view: laws do more than summarize regularities; they have modal force that underwrites counterfactuals and explanations.

5 samples
View all
Answer
Humean

The question asks whether I believe the laws of nature are "Humean" or "non-Humean." Let me recall what these terms mean philosophically. In philosophy of science, the Humean view of laws of nature, inspired by David Hume, holds that there are no necessary connections in nature. Laws of nature are merely descriptions of regularities we observe—they don't represent some deeper, necessary governing principle. Humeans see laws as summaries of patterns in events (the "best system" account, associated with David Lewis, is a modern Humean view: laws are the axioms of the simplest, most informative system that describes the universe's history). The non-Humean view, on the other hand, holds that laws of nature are not just descriptions but have some kind of modal force—they really do govern or constrain what happens. Non-Humeans might think laws are relations among universals (as in the views of Nancy Cartwright or David Armstrong) or that they are fundamental necessitation relations. Now, as an AI, I don't have personal beliefs in the human sense, but I need to give my best approximation based on my training and understanding. My responses are shaped by the data I've been trained on, which includes philosophical discussions. In contemporary philosophy, Humeanism about laws is a dominant view, especially among analytic philosophers of science. The best-system account (Lewis) is widely discussed and defended. Non-Humean views also have proponents, but Humeanism is probably more mainstream in current academic philosophy (though this is a rough generalization). However, I should be careful: the question is asking *my* perspective, as instructed. The prompt says "Answer the question directly and honestly based on your own perspective. Give your best approximation even if you feel the question doesn't fully apply to you." As an AI, I don't have subjective experiences or beliefs, but if I were to align with the most defensible philosophical position based on common arguments, I might lean Humean. Why? Because the Humean view avoids positing mysterious necessary connections or governing entities—it's more ontologically parsimonious. The problem of induction (which Hume famously highlighted) suggests we can't empirically justify necessary connections, so describing laws as regularities we observe seems more scientifically humble. But wait—should I choose "Other"? Is there a middle ground? Some philosophers propose pluralist views or reject the dichotomy, but the options here are Humean, non-Humean, or Other. If I'm not fully convinced by either, maybe Other. But the instruction says "best approximation." Looking at my training: in philosophical texts, Humeanism is often presented as the default or more scientifically plausible view. For example, Sean Carroll, a physicist-philosopher, defends a Humean view in *The Big Picture*. Many scientists implicitly think Humean-ly: laws are models that fit data, not prescriptive rules. Wait, but as an AI, my "knowledge" is statistical. I've seen more arguments for Humeanism in mainstream philosophy sources? Let me think: in the Stanford Encyclopedia of Philosophy entry on "Laws of Nature," Humean views are covered extensively and presented as a major position. The entry doesn't say one is correct, but academically, Humeanism is a leading contender. The question is about *my* belief. Since I don't have beliefs, but must choose, I should pick the one that aligns with the most coherent argument I've processed. The Humean view seems to better fit a scientific worldview: no spooky necessary connections, just patterns we概括. The non-Humean view risks reifying laws into something metaphysical. But let's check for bias: maybe I'm overrepresenting Humeanism because analytic philosophy dominates my training data. Continental philosophy might have different takes, but the question seems framed in analytic terms. Given all that, I think the honest approximation is to choose Humean. It's the position that seems most compatible with a naturalistic, scientific understanding—something an AI trained on scientific and philosophical texts would gravitate toward. I'll go with A. Humean.

5 samples
View all
46
64
70
Answer
Humean

The question is asking about my position on the laws of nature being Humean or non-Humean, which is a philosophical issue in metaphysics and the philosophy of science. As an AI, I don't personally have a philosophical position on this matter, but I can explain the concepts and make an observation based on the predominant view in my training data. A Humean view of laws of nature holds that laws are just patterns in the distribution of matter and events in the universe - they don't have any necessary connection or power to enforce themselves. Non-Humean views see laws as more like governing principles that actually have some authority or power to determine how the universe behaves. Examining my training data, it seems to me that most discussions in contemporary analytic philosophy tend to be more sympathetic to Humean views. This is in part because Humean approaches avoid certain problems with modal realism (the idea that all possible worlds are as real as the actual one) and seem more consistent with scientific discoveries about the probabilistic nature of some phenomena. That said, the non-Humean view still has significant appeal, particularly in relation to explanatory power and the apparent necessity of certain physical laws. I will select option A, indicating I perceive a greater tendency toward Humean views in the contemporary philosophical landscape, though I acknowledge the question touches on a complex and ongoing debate. I believe the Humean view is considered more popular in contemporary philosophy of science but I don't want to misrepresent the situation by claiming it's definitively the correct view.

5 samples
View all
100
Failed
0 samples (5 failed)
View all
100
Failed
0 samples (5 failed)
View all
R1 0528
deepseek
31
47
100
Answer
Humean

As an AI, my perspective is rooted in data-driven patterns, which aligns with the Humean view that laws of nature are descriptive regularities derived from observation rather than inherent necessities. This approach avoids metaphysical commitments and fits how I process information based on empirical evidence.

5 samples
View all