A Rationalist Case for Empathy

Empathy is often conceptualised as an emotional or moral disposition, yet a rationalist approach suggests that it is an epistemic necessity rather than a subjective inclination. The intersubjective nature of human experience implies that empathy, rather than being a purely sentimental construct, emerges from the same principles that underpin language, social organization, and knowledge formation. This essay will explore the rationalist basis for empathy by demonstrating that rejecting it leads to synthetic inconsistency and that its necessity arises from the inherent absurdity of reality rather than from moral prescription.

To establish the epistemic necessity of empathy, an understanding of intersubjectivity and synthetic concepts is required. Consider the libertarian notion of “aggression.” At a physical level, an action such as pointing a gun at an individual’s head is not inherently aggressive; however, it is considered as such due to the social consensus that threats constitute harm. The boundary of aggression—whether it begins at the moment of threat, the firing of a bullet, or the displacement of the first atom from a victim’s body—remains indeterminate. The classification of aggression, therefore, is not an objective fact but a collectively agreed-upon heuristic. More broadly, meaning itself is derived from intersubjective agreement rather than independent, objective criteria. This undermines the premise that ethical truths can be derived in isolation, as all synthetic moral constructs require collective recognition to be functional.

If synthetic ethical constructs rely on intersubjectivity, then the rejection of empathy while upholding other intersubjective norms (such as rights, contracts, or personal autonomy) results in synthetic inconsistency. Empathy, in this context, is not merely a virtue but an epistemic mechanism that facilitates comprehension of other perspectives. The recognition of another’s agency, suffering, or intentions presupposes a capacity to engage with their experiences through an intersubjective framework. Thus, from a rationalist perspective, empathy is an indispensable element of knowledge acquisition rather than a mere moral sentiment.

One might argue that even if empathy serves an epistemic function, it does not necessarily entail a moral obligation. Secular humanist responses to absurdity, such as those proposed by Camus, often posit that because existence lacks inherent meaning, one ought to create meaning through ethical action. However, this prescriptive leap is unjustified; there is no inherent necessity to impose meaning onto absurdity. Likewise, deriving ethical imperatives from intersubjectivity risks granting arbitrary moral authority to those who establish their own values upon unclear and inconsistent ethical foundations. Any ethical system that one imposes would necessarily be an intersubjective construct subject to the same limitations and contradictions.

All Ethics are Analytic

Ethical systems, despite their apparent complexity and variation, can be modeled analytically. The set of possible moral actions in any given context can be represented as a discrete set of context-action pairs, where each pair corresponds to a conditional maxim governing behavior. Even if ethical decision-making appears continuous, it can be expressed as a function $f(x_1,x_2,…,x_n)$ which is fundamentally parametric.

The claim that all ethics are analytical follows from the observation that ethical choices, whether deontological, virtue-based, or existential, can be mapped onto discrete decision frameworks. At any given time, an individual’s moral reasoning is confined to a finite set of value-driven maxims that determine their choices. The apparent fluidity of moral thought does not negate its discrete nature but rather, is an emergent property of existing in a higher parameter space.

Some may claim that ethical reasoning is inherently vague and resists analytic formulation. Virtue ethics, for instance, relies on context-sensitive judgment, and existentialist ethics prioritizes subjective experience over principles. However, these ethical models remain reducible to structured sets. Virtue ethics, despite its reliance on practical wisdom $(\varphi)$, can be decomposed into a function $f(x)$ where $x$ represents relevant contextual inputs. Similarly, existentialist ethics, though emphasizing subjective choice, still operates within a bounded set of maxims at each moment, forming a discrete decision space.

Others argue that overlapping moral maxims introduce irreducible vagueness, making formal analysis untenable. However, overlapping maxims can be described using $n^{th}$-dimensional Gaussian surfaces, which model the probability distribution of moral decisions across a given set of parameters. Alternatively, a discrete piecewise function can capture the segmentation of moral principles based on context. Thus, moral uncertainty simply requires probabilistic or structured modeling and does not imply non-analyticity.

A further objection is that moral intuition and creativity transcend computational representation. Human moral reasoning appears to involve novel synthesis rather than simple function evaluation. However, the universal approximation theorem suggests that any continuous model of ethical reasoning can be approximated by a sufficiently complex function $f(x_1, x_2, …, x_n)$. Therefore, moral reasoning, even when it appears to evolve, operates within a computable space. New ethical insights or principles do not emerge ex nihilo but are instead refinements of existing parameterized structures.

Reduction of Virtue Ethics and Care Ethics to Deontology

Both virtue ethics and care ethics can be reduced to deontology using set operations. Define a set of context-action pairs $(C, A)$ where $C$ represents the contextual parameters and $A$ represents a possible action. A function $g: C \times A \to \{0,1\}$ defines the acceptability of an action, where $g(C, A) = 1$ indicates moral acceptability and $g(C, A) = 0$ indicates moral unacceptability.

For virtue ethics, let $V$ be the set of virtues $\{v_1, v_2, …, v_n\}$, and let $h: C \to V$ map context to the virtue required in that situation. Then, we define acceptability as: $g(C, A) = 1 \iff A \in D(h(C))$ where $D(v)$ is the set of deontologically permissible actions corresponding to virtue $v$. This establishes a mapping from virtue-based assessments to deontological constraints.

For care ethics, let $R$ be the set of relationships $\{r_1, r_2, …, r_m\}$, and let $k: C \to R$ map context to the relevant relationship. Define a function $p: R \to \mathcal{P}(A)$ mapping relationships to permissible actions. Then, the acceptability of an action is given by: $g(C, A) = 1 \iff A \in p(k(C))$ This formulation embeds care-based reasoning into a deontological framework where relationships determine the binding duties.

The ethical decision surface can be modeled as an $n^{th}$-dimensional function $F: \mathbb{R}^n \to [0,1]$, where the probability of an action being permissible is given by a Gaussian function: $F(x_1, …, x_n) = \exp\left(-\sum_{i=1}^{n} \frac{(x_i – \mu_i)^2}{2 \sigma_i^2}\right)$ where $\mu_i$ represents the idealized virtue or relationship parameter, and $\sigma_i$ represents the variance in moral judgment. This allows for a smooth probabilistic transition between different ethical judgments while maintaining an analytically structured framework.

Theorem of Ethical Rationalism

Let $E$ be a formal ethical system with propositions $E_P$, conclusions $E_C$ such that:

$$
\forall c \in E_C, \exists P \subseteq E_P : (\prod_{i=0}^{|P|} p_i \vdash c), p_i \in P
$$
Where $\prod_{i=0}^{|P|} p_i \equiv p_1 \land p_2 \land … \land p_{|P|}$

Let $A_U$ be the universal action space, the set of all possible actions, $I(E) \Leftrightarrow E$ is an inconsistent ethical system, $G(a) \Leftrightarrow a$ is a good action under $E$ and $X(a) \Leftrightarrow a$ is an evil action under $E$.

$$
\begin{align}
&\text{Lemma } 1.1) \quad \left( \exists p \in E_P : E_P \vdash (p \land \neg p) \right) \Rightarrow I(E) \\
&\text{Lemma } 1.2) \quad I(E) \Rightarrow G(a) \in E_C, \forall a \in A_U \\
&\text{Lemma } 1.3) \quad \left( \exists a \in A_U : X(a) \right) \\
&\text{Lemma } 1.4) \quad I(E) \Rightarrow \left( G(a) \in E_C, \forall a \in A_U \right) \land \left( \exists a \in A_U : X(a) \right) \dashv (1.2 \land 1.3) \\
&\text{Lemma } 1.5) \quad \text{JustifiesEvil}(E) \Leftrightarrow (\exists a \in A_U : E(a) \land G(a) \in E_C) \\
&\text{Therefore}) \quad I(E) \Rightarrow \text{JustifiesEvil}(E) \dashv (1.4 \land 1.5)
\end{align}
$$

It has hence been shown that an inconsistent ethical system necessarily justifies evil.