(1) productive cause: the power that elicits an act of cognition (e.g., eye)
(2) objective cause: that to which the cognition tends (e.g., thing seen)
(3) formal cause: the actual exercise of the power that is able to cognize (e.g., sight)
(4) instrumental cause: the means by which the object is represented to the power (e.g., mirror).
Of these, (1)-(4) all have as their action "to make to know" (facere cognoscere); (2)-(4) 'make to know' by representing (i.e., they make something present to a power).
Against this background, John of St. Thomas builds his definition of a sign: A sign is that which represents something other than itself. The key conditions that this definition demands are (a) that there be a distinct object capable fo being manifested to a cognitive power; and (b) that there be something representative. John also wants 'represent' here to be taken in a stricter sense than it might otherwise be, so he adds three clarifying conditions, which distinguish signs from things that represent in a looser sense of the term:
(1) The sign must be better known than the signified (relative to the cognitive power involved).
(2) The sign must be (in some way) subsidiary or inferior to the thing signified.
(3) The sign must be dissimilar to the thing signified (in some relevant way).
The net effect of these conditions is to make it so that a sign is representative in that it manifests only as representative. Other things that are representative do so not only as representative, but in some looser sense. For example, a sheep does not signify another sheep qua sheep, although in a loose sense one can say that all sheep represent each other by being sheep. Likewise, God does not signify creatures, but He represents them in a looser sense in that He pre-contains their perfections in a more eminent way.
A corollary of this is that signs and images are not the same sort of thing: while an image may be a sign, not every sign is an image, and not every image is a sign. Smoke, for instance, is not an image of fire, although it is a sign of fire; and a son is not, in himself, a sign of his father, although he is certainly in himself an image of his father.
John divides signs according to two different sorts of distinction. The first distinguishes signs relative to the productive cause:
(a) Formal signs: represent without the mediation of another (e.g., concepts)
(b) Instrumental signs: represent something other than themselves by being themselves first known (e.g., footprints, words, pictures)
The second distinguishes signs relative to the objective cause:
(a) Natural signs: represent by the natures of the things they are (e.g., smoke represents fire by being its effect)
(b) Ad placitum signs: represent from the imposition of public authority (e.g., words have stable communal meanings which can be checked against authoritative references)
(c) Customary signs: represent from use alone (e.g., by sheer regularity of association, napkins on the table can represent a meal)
John notes that animals other than human beings make use of both natural and customary signs, e.g.,
We see that the animal (brutum), on seeing one thing, tends toward some other distinct thing, as when, perceiving a scent, it bounds along some path, or, hearing the roar of a lion, trembles or flees, and six hundred other cases, in which the animal does not stay with what it perceives by exterior senses, but is through it drawn to something else. Which plainly is to use a sign, that is, a representation of one thing not only for itself, but for something distinct from it. (Tractatus de signis, bk. 1, q. 6)
Likewise, animal learning shows that other animals clearly make use of customary signs.