St. Thomas explains in what sense we know the universal before the particular, and in what sense the particular before the universal:

In our knowledge there are two things to be considered.

First, that intellectual knowledge in some degree arises from sensible knowledge: and, because sense has singular and individual things for its object, and intellect has the universal for its object, it follows that our knowledge of the former comes before our knowledge of the latter.

Secondly, we must consider that our intellect proceeds from a state of potentiality to a state of actuality; and every power thus proceeding from potentiality to actuality comes first to an incomplete act, which is the medium between potentiality and actuality, before accomplishing the perfect act. The perfect act of the intellect is complete knowledge, when the object is distinctly and determinately known; whereas the incomplete act is imperfect knowledge, when the object is known indistinctly, and as it were confusedly. A thing thus imperfectly known, is known partly in act and partly in potentiality, and hence the Philosopher says (Phys. i, 1), that “what is manifest and certain is known to us at first confusedly; afterwards we know it by distinguishing its principles and elements.” Now it is evident that to know an object that comprises many things, without proper knowledge of each thing contained in it, is to know that thing confusedly. In this way we can have knowledge not only of the universal whole, which contains parts potentially, but also of the integral whole; for each whole can be known confusedly, without its parts being known. But to know distinctly what is contained in the universal whole is to know the less common, as to “animal” indistinctly is to know it as “animal”; whereas to know “animal” distinctly is know it as “rational” or “irrational animal,” that is, to know a man or a lion: therefore our intellect knows “animal” before it knows man; and the same reason holds in comparing any more universal idea with the less universal.

Moreover, as sense, like the intellect, proceeds from potentiality to act, the same order of knowledge appears in the senses. For by sense we judge of the more common before the less common, in reference both to place and time; in reference to place, when a thing is seen afar off it is seen to be a body before it is seen to be an animal; and to be an animal before it is seen to be a man, and to be a man before it seen to be Socrates or Plato; and the same is true as regards time, for a child can distinguish man from not man before he distinguishes this man from that, and therefore “children at first call men fathers, and later on distinguish each one from the others” (Phys. i, 1). The reason of this is clear: because he who knows a thing indistinctly is in a state of potentiality as regards its principle of distinction; as he who knows “genus” is in a state of potentiality as regards “difference.” Thus it is evident that indistinct knowledge is midway between potentiality and act.

We must therefore conclude that knowledge of the singular and individual is prior, as regards us, to the knowledge of the universal; as sensible knowledge is prior to intellectual knowledge. But in both sense and intellect the knowledge of the more common precedes the knowledge of the less common.

The universal is known from the particular in the sense that we learn the nature of the universal from the experience of particulars. But both in regard to the universal and in regard to the particular, our knowledge is first vague and confused, and becomes more distinct as it is perfected. In St. Thomas’s example, one can see that something is a body before noticing that it is an animal, and an animal before noticing that it is a man. The thing that might be confusing here is that the more certain knowledge is also the less perfect knowledge: looking at the thing in the distance, it is more certain that it is some kind of body, but it is more perfect to know that it is a man.

Insofar as probability theory is a formalization of degrees of belief, the same thing is found, and the same confusion can occur. Objectively, the more general claim should always be understood to be more probable, but the more specific claim, representing what would be more perfect knowledge, can seem more explanatory, and therefore might appear more likely. This false appearance is known as the conjunction fallacy. Thus for example as I continue to add to a blog post, the post might become more convincing. But in fact the chance that I am making a serious error in the post can only increase, not decrease, with every additional sentence.