# Perspective Latents as an Architectural Condition for Causal Emergence in Active Inference Agents

arXiv: https://arxiv.org/abs/2607.20708 · HTML: https://mlsanity.com/paper/2607.20708/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods.
- Builds on 7 of its 19 references that are in the map.
- First posted on arXiv 22 Jul 2026.

## Placed in

- Deep reinforcement learning methods / [Deep RL algorithms and representations](https://mlsanity.com/deeprl/deep-rl-algorithms-and-representations/)

## Builds on (most cited first, 7 of 7)

- [Integrated information theory (IIT) 4.0: Formulating the properties of phenomenal existence in physical terms](https://mlsanity.com/paper/2212.14787/) (2022)
- [Reconciling emergences: An information-theoretic approach to identify causal emergence in multivariate data](https://mlsanity.com/paper/2004.08220/) (2020)
- [Beyond integrated information: A taxonomy of information dynamics phenomena](https://mlsanity.com/paper/1909.02297/) (2019)
- [Minimal Computational Preconditions for Subjective Perspective in Artificial Agents](https://mlsanity.com/paper/2602.02902/) (2026)
- [Same World, Differently Given: History-Dependent Perceptual Reorganization in Artificial Agents](https://mlsanity.com/paper/2604.04637/) (2026)
- [The Causally Emergent Alignment Hypothesis: Causal Emergence Aligns with and Predicts Final Reward in Reinforcement Learning Agents](https://mlsanity.com/paper/2605.06746/) (2026)
- [Body-Grounded Perspective Formation and Conative Attunement in Artificial Agents](https://mlsanity.com/paper/2605.16728/) (2026)

## Abstract (arXiv)

A recent line of work measures causal emergence in reinforcement learning agents through Integrated Information Decomposition, reporting that Phir grows with training and tracks reward improvement. For active inference, this raises the question of how reward-free predictive organization relates to such information-theoretic signatures. I test this within an active inference agent whose architecture separates a fast perception latent z from a slow global latent g, where g is driven by prediction error and structurally decoupled from policy gradients. In a reward-free environmental regime-switching protocol, Phir concentrates in g; its aggregate magnitude is largely architectural and decreases with training. The substantive effect of learning becomes legible only at the atom-compositional level: decoupling flips sign from negative to positive and becomes regime-invariant under environmental change, while downward causation carries the regime-dependent adjustment. These results identify g as the architectural locus of Phir-relevant temporal organization in an active inference agent, and argue against reading scalar Phir as a direct index of learned integration.