When you write
MATLAB sees: if the number zero is less than the list of values 0, 0.4, ...
It throws an error because && is for comparing two scalars, and you've given it a scalar and a vector.
I believe what you'd like to do is compare each of the values in x to 0 (and to 1, and later to two, etc.)
There are two options, the *good* approach is to do this in a vectorized way:
y(0 <= x & x < 1) = 2.*x(0 <= x & x < 1);
y(2 <= x & x < 3) = (3/x(2 <= x & x < 3)).^(1/4);
y(3 <= x & x <= 4) = ((3/x(3 <= x & x <= 4)).^(1/4))*((x(3 <= x & x <= 4)/4).^(1/3));
Here I read this as y, for those x where 0 is less than or equal to x and x is less than 1, should be set two 2 times x for those x where 0 is less than x and x is less than 1, etc.
Note use of one & instead of two for comparing two vectors.
In general you can make this code look cleaner by naming an 'index' variable to hold the bit that goes in the parentheses.
The second approach, which will look more similar to your existing code, is to loop over the values in x:
y(i) = ((3/xx).^(1/4))*((xx/4).^(1/3));