1: \begin{definition}\label{jstabilityone} Under the setting of previous section, we define the J-weight $J_{L_2}(\scX,\scL)$ of a test configuration $(\scX,\scL)$ to equal (up to a positive multiplicative constant) the leading order term of the asymptotic Chow weight of the linear system $|L_2|$ (cf.~Theorem \ref{numericalstability}, Definition \ref{defasymChow}). Explicitly, the J-weight of a test configuration is $$J_{L_2}(\scX,\scL) = \frac{\hat{b}_0a_0 - b_0\hat{a}_0}{a_0}.$$ 

2: We define the \textit{ minimum norm} of a test configuration $(\scX,\scL)$  to be 

3:  $$\Vert (\scX,\scL)\Vert_m= J_{L}(\scX,\scL).$$

4: We say that 

5: \begin{itemize}

6:  \item $(M,L_1,L_2)$ is \emph{J-stable} (resp. J-semistable) if $$J_{L_2}(\scX,\scL)>0 \hspace{1cm}(\text{resp.}\,\,\geq0)$$  

7:  \item $(M,L_1,L_2)$ is \emph{uniformly J-stable} if for a uniform constant $\epsilon>0$  $$J_{L_2}(\scX,\scL)\geq \epsilon \Vert (\scX,\scL)\Vert_m \hspace{1.2cm}$$

8: \end{itemize}

9: for all non-trivial test configurations $(\scX,\scL)$ with normal total space.

10: \end{definition}

11: