This paper addresses a common limitation in ethical frameworks: the ability to explain structurally why ethical and unethical behavior arise, rather than only prescribing what ought to occur. Unified Natural Ethics Theory (UNET) offers a trace-sourced account of why ethical and unethical behavior occur, not merely how such behavior should be judged after the fact. UNET argues that ethics is not fundamentally an externally imposed, mind-dependent, doctrine-dependent, or deity-dependent phenomenon, but a trace-sourced emergent coherence dynamic inherent to informational propagation itself. It frames ethics as an informational, emergent, evolutive, and propagative modulation regime: a complementary constraint-pattern that biases behavior and downstream consequences toward coherence-supporting continuation under declared Boundary/Horizon (BH) conditions. Ethical adequacy is evaluated through Trace(Sequential State Trace (SST))-correspondence across retrospective -> actual -> prospective positions, under real constraints, across coupled actor-receiver relations and across scale. This manuscript preserves and extends the v0.1.1 definitional core while making several consequences of it more explicit. Coherence remains defined as (determinism + latitude), with latitude clarified as bounded variability intrinsic to determinant expression rather than as anything external to determinism. Ethical behavior is coherence; unethical behavior is dys-coherence. The manuscript also introduces the Consequential Informational Unit (CIU) as the downstream threshold at which ethically consequential process becomes behaviorally legible as a distinct responsibility-bearing unit, and it develops a more explicit architecture of dys-coherence through the Hard Wall, the distinction between mechanical failure, predatory coherence, and extractive coherence, and the role of willful or intentional unethical behavior as a cross-cutting discriminator. Meaning and Value are now carried explicitly as formal definitions within the UNET architecture, and the developmental relation between the Coordinative Sophistication Threshold (CST) and the Consequential Informational Unit (CIU) is clarified. The significance of this framework is practical as well as theoretical. It provides a structurally grounded way to distinguish ethical from unethical behavior, to identify when systems preserve themselves by collapsing or suppressing the viable latitude of the fields they inhabit, and to evaluate human, institutional, socio-technical, and Artificial Intelligence (AI) systems under a single trace-based architecture without collapsing ontology into rhetoric. In this sense, UNET is not only a theory of ethical evaluation. It is a theory of ethical emergence, ethical legibility, graduated attribution, and unethical organization across coupled fields. This paper is part of the UNET corpus. The companion law paper — Laws of Consequential Informational Dynamics (LCID) — and empirical replication record are published at OSF: https://osf.io/yjp3c/overview