The complex and hyperbolic units are usually considered two independent mathematical objects. The former underlies complex analysis, oscillation theory, and quantum mechanics, while the latter is used in hyperbolic geometry, special relativity, and algebras with Minkowski signature. At first glance, there is no direct connection between them: one satisfies the condition \(i^2=-1\), the other: \(\j^2=+1\). However, a natural question arises: are these units truly independent, or is there a more general algebraic structure in which both are different manifestations of the same mathematical object?

This note demonstrates that such a connection indeed exists. It arises after transitioning to an idempotent basis, in which the hyperbolic unit naturally decomposes into two independent complex components. This allows us to derive simple formulas for fractional powers, the logarithm, and the exponential representation of the hyperbolic unit, and also demonstrates that the complex phase may be associated not with the entire element, but only with one of its idempotent components. As a result, the complex and hyperbolic units turn out to be complementary rather than competing elements of a single four-dimensional algebraic construction.

This paper considers the relationship between the complex unit \(i\), satisfying the condition \(i^2=-1\), and the hyperbolic unit \(\j\), for which \(\j^2=+1\). The connection is constructed not through a direct identification of these units, but through an idempotent decomposition of hyperbolic algebra.

Two mutually orthogonal idempotents are introduced \[ \tag{1} \mathfrak{e} = \frac{1+\j}{2}, \qquad \bar{\mathfrak{e}} = \frac{1-\j}{2}, \] for which the following holds: ratios \[ \tag{2} \mathfrak{e}^2=\mathfrak{e}, \qquad \bar{\mathfrak{e}}^2=\bar{\mathfrak{e}}, \qquad \mathfrak{e}\bar{\mathfrak{e}}=0, \qquad \mathfrak{e}+\bar{\mathfrak{e}}=1. \]

On this basis, the fractional power of the hyperbolic unit is defined by the formula \[ \tag{3} \j^\alpha = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi\alpha}. \] Its equivalent representation in the basis \(\{1,\j\}\) is of the form \[ \tag{4} \j^\alpha = \frac12 \left[ \left(1+e^{i\pi\alpha}\right) + \j \left(1-e^{i\pi\alpha}\right) \right]. \]

The main clarification is that the logarithm of the hyperbolic unit in the extended idempotent algebra is not \(i\pi\), but \[ \tag{5} \ln\j = i\pi\bar{\mathfrak{e}} \] for the principal branch. In the general case, \[ \tag{6} \ln\j = i\pi(2k+1)\bar{\mathfrak{e}}, \qquad k\in\mathbb Z. \] Then, the following is strictly true: \[ \tag{7} e^{\ln\j} = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi} = \mathfrak{e} - \bar{\mathfrak{e}} = \j. \]

The complex unit \(i\) and the hyperbolic unit \(j\) belong to different algebraic structures: \[ \tag{8} i^2=-1, \qquad \j^2=+1. \] Therefore, the equality \[ e^{i\pi}=\j \] is not true in ordinary complex algebra. Euler's formula still yields \[ \tag{9} e^{i\pi}=-1. \]

The connection between \(i\) and \(\j\) arises only after moving to an extended algebra containing the elements \[ \tag{10} \left\{ \mathfrak{e}, \;i\mathfrak{e}, \;\bar{\mathfrak{e}}, \;i\bar{\mathfrak{e}} \right\}. \] In this algebra, the complex phase acts independently in each idempotent direction.

This approach allows us to correctly define fractional powers of the hyperbolic unit, its logarithm, and the phase structure, without identifying \(\j\) with the number \(-1\).

From definitions (1) it follows that \[ \tag{11} \mathfrak{e} + \bar{\mathfrak{e}} = 1, \] and also \[ \tag{12} \mathfrak{e} - \bar{\mathfrak{e}} = \j. \] Therefore, the unit and the hyperbolic unit have representations \[ \tag{13} 1 = \mathfrak{e} + \bar{\mathfrak{e}}, \qquad \j = \mathfrak{e} - \bar{\mathfrak{e}}. \]

The product of idempotents is zero: \[ \tag{14} \mathfrak{e}\bar{\mathfrak{e}}=0. \] Therefore, any expression of the form \[ \tag{15} Z = \mathfrak{e}A + \bar{\mathfrak{e}}B \] decomposes into two independent components. For an analytic function \(f\), the rule \[ \tag{16} f(Z) = \mathfrak{e}f(A) + \bar{\mathfrak{e}}f(B), \] holds if both functions on the right-hand side are defined.

In particular, \[ \tag{17} e^{\mathfrak{e}A+\bar{\mathfrak{e}}B} = \mathfrak{e}e^A + \bar{\mathfrak{e}}e^B, \] and for an invertible element \[ \tag{18} \ln \left( \mathfrak{e}A + \bar{\mathfrak{e}}B \right) = \mathfrak{e}\ln A + \bar{\mathfrak{e}}\ln B. \]

Since \[ \j = \mathfrak{e} - \bar{\mathfrak{e}} = \mathfrak{e}\cdot1 + \bar{\mathfrak{e}}\cdot(-1), \] it is natural to define its power componentwise: \[ \tag{19} \j^\alpha = \mathfrak{e}\,1^\alpha + \bar{\mathfrak{e}}(-1)^\alpha. \]

For the selected branch of the complex degree \[ \tag{20} (-1)^\alpha = e^{i\pi\alpha}, \] therefore \[ \tag{21} \boxed{ \j^\alpha = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi\alpha}. } \]

Substituting the definitions of idempotents, we obtain \[ \tag{22} \j^\alpha = \frac{1+\j}{2} + \frac{1-\j}{2} e^{i\pi\alpha}. \] After grouping the terms: \[ \tag{23} \boxed{ \j^\alpha = \frac12 \left[ \left(1+e^{i\pi\alpha}\right) + \j \left(1-e^{i\pi\alpha}\right) \right]. } \]

Thus, the expression in the basis \(\{1,\j\}\) is not an independent mapping, but a direct consequence of the idempotent decomposition.

For \(\alpha=0\): \[ \tag{24} \j^0 = \mathfrak{e} + \bar{\mathfrak{e}} = 1. \] For \(\alpha=1\): \[ \tag{25} \j = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi} = \mathfrak{e} - \bar{\mathfrak{e}}. \] For \(\alpha=2\): \[ \tag{26} \j^2 = \mathfrak{e} + \bar{\mathfrak{e}}e^{2i\pi} = \mathfrak{e} + \bar{\mathfrak{e}} = 1. \]

Therefore, the formula agrees with the condition \[ \tag{27} \j^2=1. \]

Since \[ e^{i\pi(\alpha+2)} = e^{i\pi\alpha}, \] we get \[ \tag{28} \j^{\alpha+2} = \j^\alpha. \] The exponent period is two.

Furthermore, \[ \tag{29} \j^{\alpha+1} = \mathfrak{e} - \bar{\mathfrak{e}}e^{i\pi\alpha}, \] and therefore \[ \tag{30} \j^{\alpha+1} = \j\,\j^\alpha. \] This confirms the usual rule for adding exponents: \[ \tag{31} \j^\alpha\j^\beta = \j^{\alpha+\beta}. \]

For \[ \alpha=\frac12 \] we have \[ e^{i\pi\alpha} = e^{i\pi/2} = i. \] Therefore, \[ \tag{32} \boxed{ \j^{1/2} = \mathfrak{e} + i\bar{\mathfrak{e}}. } \]

In the basis \(\{1,i,\j,i\j\}\), this expression has the form \[ \tag{33} \boxed{ \j^{1/2} = \frac12 \left( 1+i+\j-i\j \right). } \]

Let's check the square: \[ \tag{34} \left( \mathfrak{e} + i\bar{\mathfrak{e}} \right)^2 = \mathfrak{e} + i^2\bar{\mathfrak{e}} = \mathfrak{e} - \bar{\mathfrak{e}} = \j. \] The cross terms vanish due to the condition \[ \mathfrak{e}\bar{\mathfrak{e}}=0. \]

Consider the hyperbolic unit in idempotent form: \[ \tag{35} \j = \mathfrak{e}\cdot1 + \bar{\mathfrak{e}}\cdot(-1). \] Using rule (18), we obtain \[ \tag{36} \ln\j = \mathfrak{e}\ln1 + \bar{\mathfrak{e}}\ln(-1). \]

Since \[ \ln1=0, \] and the complex logarithm of \(-1\) has branches \[ \tag{37} \ln(-1) = i\pi(2k+1), \qquad k\in\mathbb Z, \] then \[ \tag{38} \boxed{ \ln\j = i\pi(2k+1) \bar{\mathfrak{e}}, \qquad k\in\mathbb Z. } \]

For the main branch \(k=0\): \[ \tag{39} \boxed{ \ln\j = i\pi\bar{\mathfrak{e}}. } \]

This is fundamentally different from the expression \[ \ln\j=i\pi. \] The latter equality would result in \[ e^{\ln\j} = e^{i\pi} = -1, \] rather than \(\j\). Therefore, it is incompatible with the usual definition of the exponential function in extended algebra.

Substitute the main branch: \[ \tag{40} e^{\ln\j} = e^{i\pi\bar{\mathfrak{e}}}. \] Taking into account formula (17): \[ \tag{41} e^{i\pi\bar{\mathfrak{e}}} = \mathfrak{e}e^0 + \bar{\mathfrak{e}}e^{i\pi}. \] Therefore, \[ \tag{42} e^{\ln\j} = \mathfrak{e} - \bar{\mathfrak{e}} = \j. \]

Thus, logarithm (39) is completely consistent with the definition of the hyperbolic unit.

From formula (39) it follows \[ \tag{43} \boxed{ \j = e^{i\pi\bar{\mathfrak{e}}}. } \] This equality does not mean \[ \j=e^{i\pi}. \] The complex phase \(i\pi\) acts only in the idempotent direction \(\bar{\mathfrak{e}}\), while the direction \(\mathfrak{e}\) remains unchanged.

Similarly, for arbitrary degrees: \[ \tag{44} \boxed{ \j^\alpha = e^{i\pi\alpha\bar{\mathfrak{e}}}. } \] Really, \[ \tag{45} e^{i\pi\alpha\bar{\mathfrak{e}}} = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi\alpha} = \j^\alpha. \]

In ordinary complex algebra, the argument of \(-1\) is equal to \[ \pi+2\pi k. \] However, for a hyperbolic unit, the phase structure is idempotent.

If we define the phase operator as \[ \tag{46} \Phi(\j) = -i\ln\j, \] then from formula (38) it follows \[ \tag{47} \boxed{ \Phi(\j) = \pi(2k+1) \bar{\mathfrak{e}}. } \] For the main branch: \[ \tag{48} \boxed{ \Phi(\j) = \pi\bar{\mathfrak{e}}. } \]

Therefore, the phase \(\pi\) does not correspond to the entire hyperbolic unit as a single complex object, but only to its component in the direction \(\bar{\mathfrak{e}}\).

Euler's formula \[ e^{i\pi}=-1 \] remains unchanged. The hyperbolic unit arises by replacing the number \(-1\) with a second idempotent component: \[ \tag{49} \j = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi}. \]

In other words, the hyperbolic unit consists of two independent parts: \[ \tag{50} \j = \underbrace{\mathfrak{e}}_{\text{phase }0} + \underbrace{\bar{\mathfrak{e}}e^{i\pi}}_{\text{phase }\pi}. \] The first component has zero phase, and the second is rotated by an angle of \(\pi\).

It is this two-component structure that creates the connection between the complex and hyperbolic units. This connection is expressed not by the equality \(\j=-1\), but by the expansion \[ \j = \mathfrak{e} - \bar{\mathfrak{e}}. \]

For an arbitrary element \[ \tag{51} Z = \mathfrak{e}A + \bar{\mathfrak{e}}B \] its degree is determined componentwise: \[ \tag{52} Z^\alpha = \mathfrak{e}A^\alpha + \bar{\mathfrak{e}}B^\alpha. \]

The logarithm is of the form \[ \tag{53} \ln Z = \mathfrak{e}\ln A + \bar{\mathfrak{e}}\ln B. \] The exponent: \[ \tag{54} e^Z = \mathfrak{e}e^A + \bar{\mathfrak{e}}e^B. \]

These formulas show that the extended algebra is the direct sum of two complex components. The complex unit \(i\) acts in each of them, and the hyperbolic unit \(\j\) distinguishes these components by signs.

The connection between the complex and hyperbolic units is naturally revealed through the idempotents \[ \mathfrak{e} = \frac{1+\j}{2}, \qquad \bar{\mathfrak{e}} = \frac{1-\j}{2}. \] In this basis \[ \j = \mathfrak{e} - \bar{\mathfrak{e}}. \]

The fractional power of a hyperbolic unit has the form \[ \tag{55} \boxed{ \j^\alpha = \mathfrak{e} + \bar{\mathfrak{e}}e^{i\pi\alpha}. } \] Equivalent form in the basis \(\{1,\j\}\): \[ \tag{56} \boxed{ \j^\alpha = \frac12 \left[ \left(1+e^{i\pi\alpha}\right) + \j \left(1-e^{i\pi\alpha}\right) \right]. } \]

The logarithm of the hyperbolic unit is defined by the formula \[ \tag{57} \boxed{ \ln\j = i\pi(2k+1) \bar{\mathfrak{e}}, \qquad k\in\mathbb Z. } \] For the main branch: \[ \tag{58} \boxed{ \ln\j = i\pi\bar{\mathfrak{e}}. } \]

Therefore, the exponential representation is \[ \tag{59} \boxed{ \j = e^{i\pi\bar{\mathfrak{e}}}, } \] and for an arbitrary power \[ \tag{60} \boxed{ \j^\alpha = e^{i\pi\alpha\bar{\mathfrak{e}}}. } \]

Thus, the complex phase \(\pi\) refers only to the component \(\bar{\mathfrak{e}}\), not to the entire element \(\j\). This resolves the contradiction whereby the erroneous equality \(\ln\j=i\pi\) would imply \(\j=-1\).

The resulting construction unites the complex and hyperbolic structures in a single four-dimensional basis \[ \left\{ \mathfrak{e}, \;i\mathfrak{e}, \;\bar{\mathfrak{e}}, \;i\bar{\mathfrak{e}} \right\} \] and creates a consistent foundation for further study of powers, logarithms, phases, and rotations in idempotent algebra.