Its a very silly question about a very silly circuit - not your fault, its the academic who oversimplified the theory then asked you a question that has no usable practical answer who should be blamed.
Just for laughs lets see what happens if you are actually using a 2N3904.
See the graph "Base-Emitter ON Voltage vs Collector Current" on page 3 of the
Fairchild 2N3904 datasheet. You will note that the page is labelled "Typical Characteristics", which means that considerable device to device variation is likely. To give you an idea of the possible range, on page 2 H
FE@Ic = 10 mA, Vce = 1.0 V is listed as 100 to 300. That means that for a particular Ib you will get up to a 3:1 variation in Ic between different devices. Ib and to a lesser extent Ic will certainly affect Vbe. Also note the three different lines on the graph for different temperatures, and that the voltage difference between them is about three times the difference between your supposed ON state Vbe of 0.7V and off state Vbe of 0.6V. Furthermore the graph shows Ic is typically 0.1mA @Vb=0.6V, 25°C, so that means the LED will probably be glowing some when its supposed to be off.
With that out of the way, to answer the question of what value R1 needs to be to give 0.6V on the base *ASS*U*ME*ing Ib is negligible, you simply solve R1,R2 as a potential divider.
The actual voltage will be less by Ib*Rth where Rth is the Thévenin equivalent resistance of the divider ( Rth=R1*R2/(R1+R2) ). and we can estimate an upper bound on Ib from the minimum H
FE@Ic = 0.1 mA of 40, so we can be fairly certain the difference due to Ib loading the divider wont be more than a couple of mV, provided the temperature is maintained at or below 25°C.
As this looks suspiciously like a badly posed homework question, I wont solve it for you.