It can help a little to know the science behind the blending modes but its not necessary.
http://help.adobe.com/en_US/photoshop/cs...77eba.html
The general way I consider blending modes, is to treat them as ways to paint while helping to keep your edges sharp and increase contrast. For example, multiply can be used in order to darken areas but keep the value separation the same, whereas normal mode with low opacity tends to bring values closer together, loosing your contrast and, if you used more than one stroke blur your edges. Screen is the opposite of multiply, giving you a way to brighten things. This brings us to overlay, which acts as a combination of the two, resulting in a tool to allow you to mess with darker values while leaving your lighter ones relatively unchanged.
When it comes to the dodge and burn tools, like soft light layer, I think people find themselves abusing them because it gives a way to introduce saturation variation in their colors. So they start using the tools, and in a few strokes find it looks better for some reason, then go a little crazy with them. Try to aim at using them to establish your colors in a very general sense, but that you understand to some level what color relationship you are aiming at, like this area is going to be desaturated because its a highlight, and this shadow is going to be saturated because of reflected light.
Short aside, a technical note about some layer modes that not many people seem to talk about, is a mode changing depending on whether you pick a color above 50% black or below. For example soft light, if you pick above 50%, it will 'screen or dodge', if below, it will 'multiply or burn'. So if you want to easily manage controlling your contrast around an image but keep your edges, you can use a soft brush on one soft light layer, and use black and white exactly where you want.
Understanding these kinda technical aspects about the modes is mainly to give you more flexibility in how to approach your workflow. If you only use normal mode, you will kinda always have to go big forms, and be really careful painting over your smaller forms, either constantly color picking or painting on a lower flow/opacity in order to keep the value hierarchies and gradients you made. With the layer modes, you can lose those subtle gradients as you go, but bring them back whenever you want by just using a soft brush and a specific layer mode. If you only had normal mode, you would have to paint back in details and sharpen back up edges.
So yeah, I don't think I communicated any of that very well :(. But hopefully the takeaway message is, I find that blending modes help you out by letting you keep contrast edges and detail when you want to make/change those big subtle gradients in big forms. For example, when you paint a head, you could paint an egg, then draw/paint in the features on a low opacity to make sure you keep the overall egg values. Or you could draw the egg shape, draw/paint in the features, then at the end use a layer mode to introduce that overall egg value structure.
Hope that kinda helps!
