Solution
When combining a 1D array and a 2D array, the 1D array turns into a 1xn 2D array. If the array is empty, that means the result is a 1x0 array. The transpose turns that into a 0x1 array. Combining those 2 arrays results in (1+0) for the first dimension and max(0, 1) for the second dimension. Technically this means that the first array gets padded out to the width of the second array to create the resulting 1x1 array.
According to Build Array if element is an empty array, this node produces an output array of dimension n+1 if the Concatenate Inputs option is not selected. If the Concatenate Inputs option is selected, this node produces an output array of the same dimension as the empty element array.