Definition of Tangents

When we had not learnt Calculus, we met the concept of Tangent in Circle, which was defined as the line touching the circle at ONE point.
Then, after learning Calculus, we knew that a curve could intersect with its tangent at more than one point, and a line intersecting with a curve at only one point is not necessarily a tangent. Hence, we used Limit to define tangent, which involved TWO points and we considered one approached the other to obtain the tangent.
My question is:
The definition of a tangent to a curve should be more general than that to a circle, and hence, we can say that the definition of a tangent to a circle can be derived from the definition of a tangent to a curve.
However, limit uses TWO points, even though they are very close to each other. If they overlap with each other to become ONE point, then no line occurs. So, in theory, how can we proof that the two definitions (a general curve VS a circle) are consistent?

-

That's the whole point of infinitely close stuff. Two points are so close to each other, that they actually become one. You have same issues with finding derivatives? Because, as I said, if you want to use definition of derivatives, you need to take two values - $x$ and $x+\Delta x$, where $x+\Delta x$ is infinitely close to $x$. – Kaster Feb 13 '13 at 9:25
Take a circle in the plane and a point on the circle. Choose coordinates so that the circle has equation $x^2+y^2=1$ and the point is $P=(0,1)$. Then near $P$ the equation of the circle is given by: $$y = f(x) = \sqrt{1-x^2}$$ and the tangent line has slope $f'(0) = 0$, so its equation is $y=1$ which is, indeed, a line touching the circle in a single point.