I don't think so. It is the same problem as with expand():
In [12]: ((x+cos(2*x))**2).expand(trig=1)
Out[12]:
2 2 4 2
x + 4*x*cos (x) - 2*x + 4*cos (x) - 4*cos (x) + 1
In [13]: ((x+cos(2*x))**2).expand(trig=1, multinomial=0)
Out[13]:
2
/ 2 \
\x + 2*cos (x) - 1/
Implementing it would be nice. (I honestly think that expand(trig=1) should only do trigonometric expansion, nothing else.) For expand() it is quite important that you don't expand everything possible, because it is very expensive.
Vinzent