Plato Timaeus 63

For suppose there were a solid body evenly-balanced at the center of the universe, 63a it would never be carried away to the extremities because of their uniformity in all respects; nay, even were a man to travel round it in a circle he would often call the same part of it both above and below, according as he stood now at one pole, now at the opposite. ¹ For seeing that the Whole is, as we said just now, spherical, the assertion that it has one region above and one below does not become a man of sense.
  Now the origin of these names and their true meaning which accounts for our habit of making these verbal distinctions even about the whole Heaven, 63b we must determine on the basis of the following principles. Suppose that a man were to take his stand in that region of the Universe in which the substance of fire has its special abode, and where also that substance to which it flies is collected in largest bulk; and suppose that, having the power to do so, he were to separate portions of the fire and weigh them, putting them on scales and lifting the balance and pulling the fire by force into the dissimilar air, it is obvious that he will force the smaller mass more easily than the larger. 63c For if two masses are lifted up simultaneously by a single effort, the smaller will necessarily yield more and the larger less, owing to its resistance, to the force exerted; and the large mass will be said to be heavy and moving down, the small light and moving up. Now this is just what we ought to detect ourselves doing in our region here. Standing on the earth and detaching various earthy substances, and sometimes pure earth, we pull them into the dissimilar air by force and against nature, since both these kinds cleave to their own kindred; 63d and the smaller mass yields more easily, and follows first, as we force it into the dissimilar kind; wherefore we name it light, and the region to which we force it above; and the conditions opposite thereto we name heavy and below. Thus, these must necessarily differ in their mutual relations, because the main masses of the Kinds occupy regions opposite to one another; for when we compare what is light in one region with what is light in the opposite region, and the heavy with the heavy, the below with the below, 63e and the” above with the above, we shall discover that these all become and are opposite and oblique and in every way different in their mutual relations. There is, however, this one fact to be noticed about them all, that it is the passage of each kind to its kindred mass which makes the moving body heavy, and the region to which such a body moves below; while the opposite conditions produce the contrary results. ² Let this, then, stand as our account of the causes of these conditions.