This is from Exercise 26, Chapter 1, in Stein and Shakarchi's Functional Analysis.
Suppose $1 < p_0, p_1 < \infty$ and $1/p_0+ 1/q_0 = 1$ and $1/p_1 + 1/q_1 = 1$. Show that the Banach spaces $L^{p_0} \cap L^{p_1}$ and $L^{q_0} + L^{q_1}$ are duals of each other up to an equivalence of norms.
Below are the definitions of $L^{p_0} \cap L^{p_1}$ and $L^{p_0} + L^{p_1}$.
Define the norm of $f \in L^{p_0} \cap L^{p_1}$ as $$\|f\|_{L^{p_0} \cap L^{p_1}} = \|f\|_{L^{p_0}} + \|f\|_{L^{p_1}}.$$
$L^{p_0}+L^{p_1}$ is defined as the vector space of measurable functions $f$ on a measure space $X$, that can be written as a sum $f=f_0+f_1$ with $f_0\in L^{p_0}$ and $f_1\in L^{p_1}$. Define $$\|f\|_{L^{p_0}+L^{p_1}}=\inf\big\{\|f_0\|_{L^{p_0}}+\|f_1\|_{L^{p_1}}\big\},$$ where the infimum is taken over all decomposition $f=f_0 + f_1$ with $f_0\in L^{p_0}$ and $f_1\in L^{p_1}$.
What is meant by "dual space" is as follows.
For every bounded linear functional $l$ on $L^{p_0}+L^{p_1}$ there is a unique $g \in L^{q_0} \cap L^{q_1}$ so that $$l(f) = \int_X f(x)g(x) d\mu(x), \quad \text{for all $f \in L^{p_0}+L^{p_1}$}$$
Moreover, $\|l\|_{(L^{p_0}+L^{p_1})*} = \| g \|_{L^{p_0} \cap L^{p_1}}$.
By modifying Lemma 4.2 in the textbook, it is not too hard to prove the case above, i.e., $L^{p_0} \cap L^{p_1}$ is the dual space of $L^{p_0} + L^{p_1}$. I have trouble to prove the opposite, that $L^{p_0} + L^{p_1}$ is the dual space of $L^{p_0} \cap L^{p_1}$.
Below is my attempt following the hint provided by @PhoemueX.
Edited following PhoemueX's and Oliver Díaz's comments.
We first define the subspace $$S \equiv \{(f,f):f \in X \cap Y\}$$ of $X \times Y$, where $X=L^{p_0}$ and $X=L^{p_1}$
For any linear functional $l$ on $X\cap Y$, we can define a linear functional on $S$ $$\lambda_0\left((f_0,f_1)\right) = l\left(\frac{f_0+f_1}{2}\right),$$ and a sub-linear function $p$ on $X \times Y$ $$p\left((f_0,f_1)\right) = \|l\|_{(X\cap Y)^*} \frac{\|f_0\|_X + \|f_1\|_Y}{2}.$$
On $S$, we have $\lambda_0\left((f_0,f_1)\right) \leq p\left((f_0,f_1)\right)$. So we can use the Hahn-Banach theorem to extend $\lambda_0$ to $X \times Y$.
It is easy to show that the dual of $X \times Y$, with the norm definition $\left\|(f_0,f_1)\right\|_{X \times Y} = \|f_0\|_X + \|f_1\|_Y$ is simply $X^* \times Y^*$, with the norm definition $\left\|(g_0,g_1)\right\|_{X^* \times Y^*} = \max\big(\|g_0\|_{X^*}, \|g_1\|_{Y^*}\big)$. (There are many other norm definitions possible. This one is simply chosen for the original problem.)
So that we can find $(g_0,g_1) \in X^* \times Y^*$, and have $$\lambda\left((f_0,f_1)\right) = \int_X \bigg(f_0(x)g_0(x) + f_1(x)g_1(x) \bigg) d\mu(x).$$
Setting $f_0 = f_1 = f \in X \cap Y$ and $g = g_1 + g_2 \in L^{q_0} + L^{q_1}$, we get the desired the result.
EDIT: I have found another proof that requires heavier machinery (i.e. Orlicz spaces) but gives us a more general result. I will post it one another day.