Aristotle, in Topica 1.IV, discussed the basis of propositions and problems, the later of which are merely the former recast in question form.  A proposition is assumed to be of the form, “A is an X.”  Thus, we have the traditionally assigned term, “predicable,” that is, something which can be predicated of a thing.  For example, “John is a two footed animal” or “Man is capable of learning to type” are two propositions and each subject, “John” and “Man” has something predicated of it, respectively a “two footed animal,” and “capable of learning to type.”  Whether Aristotle has intended to exhaustively account for all predicables, that is, attributes which are said to be about some thing “X”, we do not know.  It does appear though, that this four-fold division into definition, property, genus and accident is able to suffice for all practical purposes.
Let us look now at all the first predicable, definition.

One must speak about what definition is, what a property is, what a genus is, and what an accident is.  A definition is a phrase which signifies the essence.  A phrase is either given in exchange for a word or a phrase in exchange for another phrase.  For it is possible for one of the things being signified to be defined by a phrase.  When people make a proposition by assigning a name by some way, it is obvious that they do not give the definition of a thing, since every definition is a kind of phrase.  Indeed one must consider such a thing “definitional”, such as, “the seemly is noble.”  And likewise, “Whether perception and knowledge are the same thing or different?”  For most of the effort concerning definitions is whether something is the same or different.  Simply let all those things be called “definitional” which come under the same method as definitions.  That all the things just said are such, it is obvious from these.  When we are able to argue that things are the same or different, we are also able to undertake this same way regarding definitions.  For pointing out that things are not the same, we will have destroyed the definition.  The thing just said cannot be said in reverse though.  For it is not sufficient for proving a definition to point out that something is the same; though it is sufficient for demolishing a definition to point out that something is not the same.[i]  Topica 101b36-102a17 

We might have expected Aristotle to have given us a clear and detailed definition of “definition,” since it this idea, at least from the time of Socrates, that appears to be the starting point for philosophical discourse.  Aristotle is stinting as often, however, and we do not quite receive as full an explanation as we had wished.

Aristotle begins by pointing out that a definition is necessary and sufficient to describe the essence of something.  Because it sets up the minimal conditions of what something is, a definition cannot be as simple as a single word.  A definition must be a phrase.  For example the phrase, “A flat-surfaced, 4 legged object,” could correspond, as a definition, to the word, “table.”  Likewise, “[A table is] a flat-surfaced, 4 legged object,” could be used in place of another phrase, such as, “horizontally raised plane.”  We could alternatively, for instance, say, “a horizontally raised plane” is a “flat-surfaced, 4 legged object.”

When something is merely a copulative pairing, such as, “x is y,” Aristotle prefers to call it “definitional” (horikos) rather than properly a definition (horos/horismos).

Aristotle makes the fascinating point that much of the time and effort in coming up with definitions involves determining similarities and differences between two things.  Furthermore, by giving either an example or a counterexample I can prove or disprove a definition.  If I maintain that a, “horizontally raised plane” is, in fact, a “table,” such a designation may be insufficiently determinative of the “essence” of the thing in question.  Thus the phrase, “horizontally raised plane” is insufficient for demonstrating a given definition, at least by the method of providing examples.  Another way of saying this is that if I point out objects which are really tables, they would also match the description of “horizontally raised plane.”  That an actual table happens to match up with my definition of table might in fact be a mere coincidence, owing to the broad and accommodating definition I have offered.  As Aristotle correctly puts it, therefore, you cannot prove a definition in this way.  On the other hand, however, one could disprove a definition by offering a counterexample.  “Horizontally raised plane” also includes things such as desks, cabinets or even geographical landmarks such as plateaus.  Because of the imprecision in this case (perhaps flat-out error in another example of a definition), the definition of a table as a, “horizontally raised plane” has been refuted.  Thus counterexamples have proved the definition incorrect.