You are, whether you know it or not, raising the issue of science methodology. The first step in any investigation of phenomona is to categorize that phenomenon. You start by defining the boundaries—in this case, who you prefer to have sex with. Then you look at human sexual behavior and see all the range of sexual behavior and at some point, you try to group sexual behavior into groups that make sense based on the way our minds work.
All category schemas are arbitrary. I always like to use the example of flora. Someone, at some point, had to come along at decide that there are trees, shrubs and grasses. There will always be examples of flora that seem to fit in more than one category. You have to draw a line, and that line is always arbitrary.
For example, look at race. There are so many hues of skin color, and yet people seem to have a pretty good idea of what is black and what is white. And yet, there are yellow and other hues in between that different people will assign to different races. Race is assigned in two ways: self-chosen and imposed by others. How you see yourself may be different from how I see you. My point is that these category boundaries are often quite arbitrary. There is no effort to define the boundaries of the categories very precisely. But even if they are defined well, there will always be exceptions that make the definitions look stupid.
The simplest form of categorization is, as @the100thmonkey pointed out, the dialectic: either, or. One or zero. It is or it isn’t. This, in science, is quite unsophisticated. But it is easy for people to see.
Next, we can have a number of categories. Typically, when we allow more than one category, there are between two and nine categories with seven being the most common number of categories. The thought is that it is seven because research shows that that is about the most number of differences people can keep in mind at the same time.
Once we reach seven, we tend to subcategorize rather than expanding the number of top level categories.
This is all to say that the number of categories of sexual behavior is more dependent on how we think than on what is really out there. In fact, there are people who think there are more that two categories of sexual behavior. Maybe they think there are seven categories. Ask SimonedeBeauvoir. She would know.
So if we didn’t have homosexuality as one of the categories, then there would be something else to define our either/or understanding. There will never be one category. It will always be, at the very least, this or not-this. For more complex thinkers, there will also be a number of not-thats.
So yes, you can think of it as “opposites,” although they aren’t opposites. You should think of it as this and not-this. Or that and non-that. It is a way of thinking, I believe, that is built into the architecture of our brains. It got there because it is useful. It helps of make sense of our world. It helps us survive.
If you think about it, it will probably make sense to you. If we do not differentiate between things, then we are all one. When the tiger leaps out at us, we think it is us, and soon we are tiger food. We have to distinguish things if we are to survive. Often times we have not made distinctions and it has killed those of us who didn’t make the correct distinction. A good way of learning. Harsh, but effective.
In addition, we may also make distinctions where it is not so important. It’s hard to tell. How important is it to distinguish based on sexual preferences? I don’t know. Is there any harm in sexual preferences? Or is it all good? I don’t see the harm. So I’m not sure it matters except as an easy way to determine who you want to hang out with.
[Edit]
To answer prolificus amendment to the question: if there were not what he calls “opposites,” we would be dead.