= Solution
Each annulus has a <torque> proportional to $\mathbf J_h\times\delta\mathbf J_d$, so its scalar product with $\mathbf J_h$ is zero. Summing over annuli preserves this property even when the proportionality coefficient varies with radius. Therefore $\boxed{\mathbf T\cdot\mathbf J_h=0}$ for the total <torque> due to <Lense-Thirring precession>.
The <torque> specified is the <torque> on the disk. Assuming a closed disk-plus-hole angular-momentum budget with no external <torque>, its reaction on the <black hole> has the opposite sign:
$$
\boxed{\dot{\mathbf J}_h=-\mathbf T
=-K_1\mathbf J_h\times\mathbf J_d
-K_2\mathbf J_h\times(\mathbf J_h\times\mathbf J_d),\qquad
\dot{\mathbf J}_d=\mathbf T.}
$$
Thus $\dot{\mathbf J}_t=0$. The $K_1$ term describes <precession>, because it is perpendicular to both spins. Since $\mathbf J_h\times\mathbf J_d=\mathbf J_h\times\mathbf J_t$, the hole precesses about the fixed total vector. The $K_2$ term changes the spin directions dissipatively. Writing $\mathbf J_{d,\perp}$ for the disk component perpendicular to the hole, the <vector triple product identity> gives
$$
\mathbf J_h\times(\mathbf J_h\times\mathbf J_d)=-J_h^2\mathbf J_{d,\perp}.
$$
A positive $K_2$ removes this component from the disk and transfers it into the hole's changing direction. This is <black-hole and disk alignment under internal torque>.
Since $\mathbf T$ is perpendicular to $\mathbf J_h$,
$$
\frac d{dt}J_h^2=2\mathbf J_h\cdot\dot{\mathbf J}_h=-2\mathbf J_h\cdot\mathbf T=0.
$$
Hence $\boxed{J_h\text{ is constant}}$, even if the <torque> coefficients vary in time.
Now differentiate the scalar product with the conserved total vector. The <precession> term contributes zero, and the <vector triple product identity> gives
$$
\begin{aligned}
\frac d{dt}(\mathbf J_h\cdot\mathbf J_t)
&=-\mathbf T\cdot(\mathbf J_h+\mathbf J_d)=-\mathbf T\cdot\mathbf J_d\\
&=K_2\left[J_h^2J_d^2-(\mathbf J_h\cdot\mathbf J_d)^2\right]\\
&=\boxed{K_2|\mathbf J_h\times\mathbf J_d|^2=A.}
\end{aligned}
$$
Also $\mathbf J_h\cdot\mathbf J_t=J_h^2+\mathbf J_h\cdot\mathbf J_d$, so constancy of $J_h$ gives
$$
\boxed{\frac d{dt}(\mathbf J_h\cdot\mathbf J_t)
=\frac d{dt}(\mathbf J_h\cdot\mathbf J_d).}
$$
Finally, $J_t^2=J_h^2+J_d^2+2\mathbf J_h\cdot\mathbf J_d$ is constant. Differentiating proves
$$
\boxed{\frac d{dt}J_d^2=-2A.}
$$
Thus the disk's spin magnitude decreases while the hole's magnitude is fixed; their vector sum remains fixed as well.
For the alignment conclusion, assume $J_h,J_t>0$ and let $c=\cos\theta=\mathbf J_h\cdot\mathbf J_t/(J_hJ_t)$. Since $\mathbf J_h\times\mathbf J_d=\mathbf J_h\times\mathbf J_t$,
$$
\boxed{\dot c=K_2J_hJ_t(1-c^2).}
$$
For a noncollinear initial state, $|c(0)|<1$, this integrates exactly to
$$
\operatorname{artanh}c(t)=\operatorname{artanh}c(0)+J_hJ_t\int_0^tK_2(s)\,ds.
$$
Consequently a fixed positive $K_2$, or more generally a positive coefficient of divergent integrated strength, gives $c(t)\to1$. In this persistent dissipative evolution, $\boxed{\mathbf J_h\to J_h\widehat{\mathbf J}_t}$: the hole aligns asymptotically with the total <angular momentum>. The exactly antiparallel configuration $c=-1$ is a stationary, nongeneric exception; when $J_t=0$ there is no total-vector direction to align with.
If the phrase $K_2>0$ allows arbitrary time dependence, positivity alone does not logically force complete eventual alignment. The <persistent-torque criterion for asymptotic spin alignment> is the needed qualification. For example, take $J_h=J_t=1$, $K_1=0$, $K_2(t)=e^{-t}$ and $c(0)=0$. The same <torque> equations give $c(t)=\tanh(1-e^{-t})$, with limit $\tanh1<1$. The final configurations below use the intended persistent-<torque> regime, including constant positive $K_2$.
Once the hole has aligned, <conservation of angular momentum> fixes the disk without any additional <torque> calculation:
$$
\boxed{\mathbf J_{h,f}=J_h\widehat{\mathbf J}_t,\qquad
\mathbf J_{d,f}=(J_t-J_h)\widehat{\mathbf J}_t.}
$$
The disk and hole therefore finish parallel only if $J_t>J_h$. If $J_t<J_h$, they finish antiparallel: this is <counteralignment of a black hole and accretion disk>. In terms of the initial disk-hole angle $\theta_0$, squaring $J_t$ shows that counteralignment occurs when
$$
\boxed{\cos\theta_0<-\frac{J_{d,0}}{2J_h}.}
$$
If $J_t=J_h$, the final disk <angular momentum> vanishes and has no spin direction. Thus alignment of the hole with the total vector does not imply alignment of both component spins.
Back to article page
